Switch branches/tags
Find file Copy path
5086 lines (4317 sloc) 270 KB

Datadog Agent 6 has been officially released and the release notes can be found here. Agent 5 will continue to see releases.


5.26.0 / 08-01-2018

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.26.0 tag for the list of changes on the core checks.

Please refer to the Trace Agent 6.4.1 tag for more information on the Trace Agent.

Please refer to the Process Agent 5.26.0 tag for more information on the Process Agent.

5.25.3 / 07-17-2018

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.25.3 tag for the list of changes on the core checks.

Please refer to the Trace Agent 6.3.3 tag for more information on the Trace Agent.

Please refer to the Process Agent 5.25.3 tag for more information on the Process Agent.

5.25.2 / 07-11-2018

Linux, Windows, Docker and Source Install


Dependency Upgrades

  • [SECURITY] The agent configuration loader has been patched to increase the security of loaded configuration files.


Please refer to the Integrations Core Checks 5.25.2 tag for the list of changes on the core checks.

5.25.1 / 06-28-2018

Linux, Windows, Docker and Source Install


Dependency Upgrades


Please refer to the Integrations Core Checks 6.3.1 tag for the list of changes on the core checks.

Please refer to the Trace Agent 6.3.1 tag for more information on the Trace Agent.

Please refer to the Process Agent 6.3.1 tag for more information on the Process Agent.

5.25.0 / 06-20-2018

Linux, Windows, Docker and Source Install


Dependency Upgrades


Please refer to the Integrations Core Checks 6.3.0 tag for the list of changes on the core checks.

Please refer to the Trace Agent 6.3.0 tag for more information on the Trace Agent.

Please refer to the Process Agent 6.3.0 tag for more information on the Process Agent.

  • [BUGFIX] JMXFetch: make classpath separator adjust to the OS. See #3729 (Thanks @jvanlieshout)
  • [BUGFIX] Source install: fix integrations base package name change + bashisms. See #3738
  • [BUGFIX] IO: iostat - fix for newer iostat output. See #3733
  • [BUGFIX] IO: iostat - fix for broken %util metric due to missing cast. See #3733
  • [IMPROVEMENT] IO: iostat - capping values to 100%. See #3727
  • [IMPROVEMENT] IO: iostat - adding test for cap logic. See #3741
  • [IMPROVEMENT] Core: add host tag for docker swarm node role. See #3735 (Thanks @devonboyer)
  • [IMPROVEMENT] Core: warning message to clarify check_rate option available. See #3727
  • [SANITY] Linux: Reintroduced system stats tests. See #3733
  • [SANITY] Packaging/Build: Explicilty pin urllib to avoid inadverted updated. See #3746
  • [OTHER] APM enabled by default. See #3749
  • [OTHER] Our embedded python unicode interpreter is now UCS4. See omnibus-software-191

5.24.0 / 05-11-2018

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 6.2.0 tag for the list of changes on the core checks.

Please refer to the Trace Agent 6.2.0 tag for more information on the Trace Agent.

Please refer to the Process Agent 6.2.0 tag for more information on the Process Agent.

  • [IMPROVEMENT] Windows: Add pdh data to flare. See #3719
  • [IMPROVEMENT] JMXFetch: Bump to 0.20.0. See #3726 and jmxfetch 0.20.0 tag
  • [OTHER] subprocess_output: Add compatibility layer with datadog_checks_base. See #3723
  • [OTHER] Windows: Use PDH libs provided in datadog_checks_base. See #3691

5.23.0 / 03-21-2018

Linux, Windows, Docker



Please refer to the Integrations Core Checks 5.23.0 tag for the list of changes on the core checks.

Please refer to the Process Agent 5.23.0 tag for more information on the Process Agent.

  • [BUGFIX] Windows: Fix pdh counters on localized versions of windows. See #3713
  • [BUGFIX] Remove API key if included in processes metadata payload. See #3712
  • [BUGFIX] Avoid failing the configcheck command if the instances is empty. See #3704
  • [BUGFIX] Kuernetes: fix logic to collect ownerreference for k8 1.9. See #3667
  • [BUGFIX] Fix an error when the agent failed to find the name of a docker image. See #3669
  • [BUGFIX] pause-container: update the list of the pause containers. See #3670
  • [IMPROVEMENT] jmxfetch: include version 0.19.0. See #3711
  • [IMPROVEMENT] Logs an INFO message instead of an ERROR when the host is deemed as Dockerless. See #3657
  • [IMPROVEMENT] Windows: change default datatype for pdh queries to double. See #3703
  • [OTHER] Remove the deprecated option skip_event in network checks. See #3643

5.22.3 / 03-08-2018

Linux, Windows, Docker


This release only ships changes to the process-agent.


Please refer to the Process Agent 5.22.3 tag for more information on the Process Agent.

5.22.1 / 02-26-2018

Linux, Windows, Docker


This release only ships changes to the process-agent.


Please refer to the Process Agent 5.22.1 tag for more information on the Process Agent.

5.22.0 / 02-13-2018

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.22.0 tag for the list of changes on the core checks.

Please refer to the Process Agent 5.22.0 tag for more information on the Process Agent.

Datadog will be rotating the GPG key used to sign our APT repositories on Feb 28 2018. More details and the new key can be found in our knowledge base.

  • [IMPROVEMENT] Collect total system memory for Mac OS X. See #3169
  • [IMPROVEMENT] JMXFETCH: include version 0.18.1. See #3543
  • [BUGFIX] Fix prometheus check summary and histogram. See #3617
  • [IMPROVEMENT] Prometheus Check: add generic label joins and node to hostname. See #3623
  • [BUGFIX] Fixes issue with auto-discovery and wheels. See #3629
  • [IMPROVEMENT] Add "process_agent_enabled" to example datadog.conf. See #3630
  • [BUGFIX] Fix issue with creating tags from conf in IIS. See #3635
  • [IMPROVEMENT] Prometheus Check: add SSL support. See #3640
  • [IMPROVEMENT] Deduplicate tags on metrics, events and service_checks. See #3641
  • [BUGFIX] Fix "no_proxy" option when using requests. See #3644 and #3645
  • [IMPROVEMENT] Adding "jmx_custom_jars" option to datadog.yaml. See #3648
  • [IMPROVEMENT] Windows: the agent service is no longer responsible for starting the APM service. See #3655
  • [DEPRECATE] Bundled "checks.d" is now a deprecated location as checks are shipped as wheels. See #3661
  • [IMPROVEMENT] Windows: Start subservices (trace/process) if they're configured on service start. See #3663
  • [IMPROVEMENT] Extend pause-container list. See #3674

5.21.2 / 01-31-2018

Linux, Windows, Docker


This release only ships changes to the process-agent.


Please refer to the Process Agent 5.21.2 tag for more information on the Process Agent.

5.21.1 / 01-31-2018

Linux, Windows, Docker


This release only ships changes to the trace-agent.


Please refer to the Trace Agent 5.21.1 tag for the list of changes on the Trace Agent.

5.21.0 / 01-10-2018

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.21.0 tag for the list of changes on the core checks.

Please refer to the Process Agent 5.21.0 tag for more information on the Process Agent.

  • [BUGFIX] Configuration: Use : instead of = in APM config. See #3599 (Thanks @brianwolfe)
  • [BUGFIX] PDH: Return with metric list empty if can't connect to host. See #3612
  • [IMPROVEMENT] Source install: Add host tagging to See #3551 (Thanks @iancward)
  • [IMPROVEMENT] Kubernetes: get nodeName and nodeIP even if the agent isn't inside a Pod. See #3614
  • [IMPROVEMENT] Docker: add flag to image_name_extractor to return short image name. See #3622
  • [IMPROVEMENT] WMI: Skip properties in WMI query results if not in arguments. See #3618
  • [IMPROVEMENT] Log metadata payload as debug to be added to flares. See #3625
  • [IMPROVEMENT] Don't spam logs with exceptions when checks aren't wheels. See #3620
  • [IMPROVEMENT] bump protobuf to 3.5.1. See #3619
  • [IMPROVEMENT] Aggregator: ignore all dist metrics (metric type starting with 'd'). See #3611
  • [IMPROVEMENT] Auto-Discovery: add kube container name in service discovery containers. See #3609
  • [IMPROVEMENT] Windows: change restart to stop/start with deps. See #3565
  • [FEATURE] Allow running checks packaged as wheels. See #3555
  • [FEATURE] Filehandles: include allocated, allocated but unused, used, and max filehandler metrics. See #3608 (Thanks @kennydo)
  • [FEATURE] Adds an option to disable IPv6. See #3603
  • [FEATURE] Windows: update pdh library to support generic pdh based checks. See #3581
  • [FEATURE] Windows: update base libraries for PDH based checks. See #3607
  • [FEATURE] Collect OpenStack metadata and tags community. See #3411 (Thanks @mtougeron)

5.20.2 / 12-06-2017

Linux, Docker and Source Install



Please refer to the Integrations Core Checks 5.20.2 tag for the list of changes on the core checks.

Please refer to the Process Agent 5.20.2 tag for more information on the Process Agent.

5.20.1 / 11-29-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.20.1 tag for the list of changes on the core checks.

Please refer to the Trace Agent 5.20.0 tag for the list of changes on the Trace Agent.

Please refer to the Process Agent 5.20.1 tag for more information on the Process Agent.

  • [BUGFIX] Windows: fixing flakes, stubbing WinPDHCounter #3596

5.20.0 / 11-21-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.20.0 tag for the list of changes on the core checks.

Please refer to the Trace Agent 5.20.0 tag for the list of changes on the Trace Agent.

Please refer to the Process Agent 5.20.0 tag for more information on the Process Agent.

  • [BUGFIX] Audo Configuration: Fix AD port parsing. See #3553
  • [BUGFIX] Core: Fixes normalization of metric names, removing @. See #3529
  • [BUGFIX] Core: Only start trace agent on windows if its enabld. See #3554
  • [BUGFIX] Kubernetes: fix tagging on k8s 1.8+ clusters. See #3563, #3586
  • [BUGFIX] System Checks: Fixes grabbing of file handles. See #3539 (Thanks @matlockx)
  • [BUGFIX] Audo Configuration: Fix enabling condition. See #3589
  • [IMPROVEMENT] Core: Won't start if API Key is too short. See #3546
  • [IMPROVEMENT] Core: Won't start agent on windows if API Key invalid. See #3547
  • [IMPROVEMENT] Kubernetes: Adds more tags to events. See #3548
  • [IMPROVEMENT] Core: Decrease log noise at startup. See #3547
  • [IMPROVEMENT] Kubernetes: Filter kubernetes events based on timestamp. See #3549
  • [IMPROVEMENT] Audo Configuration: Adds docker labels as tags. See #3564
  • [IMPROVEMENT] Containers: Move most container tags to metadata. See #3557
  • [IMPROVEMENT] Kubernetes: Allow overriding the pod name. See #3571 (Thanks @darnaut)
  • [IMPROVEMENT] Windows: Remove check for pidfile. See #3574
  • [IMPROVEMENT] Windows: Change restart to start/stop with deps. See #3565
  • [FEATURE] Windows: Add a PDH base check. See #3581

5.19.0 / 11-03-2017

Linux, Windows, Docker and Source Install

Windows released on 11-03-2017

Linux, Docker and Source Install released on 11-08-2017


This release includes:

  • the trace-agent on Windows, as a Beta feature, and related changes on the Windows service management
  • features and bugfixes for the trace-agent and process-agent on Linux, Docker and Source Install - please check their respective changelogs for more details
  • changes on Autodiscovery & container support
  • bugfixes on some core checks


Please refer to the Integrations Core Checks 5.19.0 tag for the list of changes on the core checks.

Please refer to the Trace Agent 5.19.0 tag for the list of changes on the Trace Agent.

Please refer to the Process Agent 5.19.0 tag for more information on the Process Agent.

  • [FEATURE] Autodiscovery: Add docker labels as tags in autodiscovery. See #3564
  • [FEATURE] Windows: Enable the apm service based on the configuration file. See #3554
  • [FEATURE] Windows: Include the trace agent as part of the Datadog Agent. See dd-agent-omnibus-203 & dd-agent-omnibus-204
  • [IMPROVEMENT] Switch most of container host tags to host metadata. See #3557
  • [IMPROVEMENT] Windows: Tighten permissions on datadog.conf to only be readable by admins and local system. See dd-agent-omnibus-202

5.18.1 / 10-13-2017

Linux, Docker and Source Install


This release includes bugfixes for the trace-agent and process-agent only - please check their respective changelogs for more details.


Please refer to the Trace Agent 5.18.1 tag for the list of changes on the Trace Agent. Please refer to the Process Agent 5.18.1 tag for more information on the Process Agent.

5.18.0 / 10-10-2017

Linux, Windows, Docker and Source Install


Dependency Changes


Please refer to the Integrations Core Checks 5.18.0 tag for the list of changes on the core checks. Please refer to the Trace Agent 5.18.0 tag for the list of changes on the Trace Agent. Please refer to the Process Agent 5.18.0 tag for more information on the Process Agent.

  • [FEATURE] JMXFetch: add canonical rate. See #3494, jmxfetch-156 (Thanks @arrawatia)
  • [IMPROVEMENT] Core: adding docker inspect to flare. See #3536
  • [IMPROVEMENT] Windows: enhanced PDH counter support. See #3536
  • [BUGFIX] RHEL/CentOS7[systemd]: enable datadog-agent service on post-install. See dd-agent-omnibus-201
  • [BUGFIX] Windows: fix PDH counter i18n issues. See #3536
  • [BUGFIX] Kubernetes: fix image lookup. See #3532
  • [BUGFIX] Mesos: do not crash if ENV in CONF is none. See #3528
  • [SANITY] WMI: renamed library from wmi_check to winwmi_check. See #3521
  • [SECURITY] Core: Redact sensistive information in config dumps. See #3534
  • [SECURITY] Core: fix potential leak of API key into collector/dogstatsd logs. See #3537
  • [DEPRECATE] deprecate read_config. See #3508
  • [DEPRECATE] Docker: remove access to deprecated top-level NetworkSettings. See #3538

5.17.2 / 09-18-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.17.2 tag for the list of changes on the core checks. Please refer to the Process Agent 5.17.2 tag for more information on the Process Agent.

Dependency changes


  • [REGRESSION] Kubernetes: partially revert _locate_kubelet from #3497. See #3523

5.17.1 / 09-12-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.17.1 tag for the list of changes on the core checks. Please refer to the Trace Agent 5.17.1 tag for the list of changes on the Trace Agent. Please refer to the Process Agent 5.17.1 tag for more information on the Process Agent.

  • [IMPROVEMENT] Autodiscovery: Make Autodiscovery wait on docker/kubelet. See #3497
  • [IMPROVEMENT] Docker: catch docker.errors.NotFound for nonexisting images. See #3507
  • [BUGFIX] Autodiscovery: Rework Autodiscovery image name extractor to support new swarm format. See #3511
  • [BUGFIX] Dogstreams: Retry reading files that could't be read on agent startup. See #2815, thanks @jalaziz
  • [BUGFIX] Flare: Update api key regex to support more formats. See #3509
  • [BUGFIX] Kubernetes: Add image_name_resolver to kubeutil wrapping dockerutil. See #3505
  • [BUGFIX] Kubernetes: Fix leader election's lease parameter and node name. See #3506

5.17.0 / 08-28-2017

Linux, Windows, Docker and Source Install


Dependency changes


Please refer to the Integrations Core Checks 5.17.0 tag for the list of changes on the core checks. Please refer to the Trace Agent 5.17.0 tag for the list of changes on the Trace Agent. Please refer to the Process Agent 5.17.0 tag for more information on the Process Agent.

  • [FEATURE] Autodiscovery: Allow autoconf templates in docker labels. See #3451
  • [FEATURE] JMXFetch: Add list_jvms command on dd-agent. See #2962
  • [IMPROVEMENT] Azure: Use VM ID as a host alias. See #3402
  • [IMPROVEMENT] Collector: Split check runs from main agent payload. See #3446
  • [IMPROVEMENT] Collector: Split metric payloads larger than 2MB into smaller payloads. See #3454
  • [IMPROVEMENT] Collector: On info page, display unknown version if module load fail, custom if custom check. See #3488
  • [IMPROVEMENT] Debian packaging: Add gnupg as dependency, and suggest dirmngr. See #3492 and dd-agent-omnibus-191
  • [IMPROVEMENT] Docker: Add openshift/origin-pod to default docker ignore list. See #3460
  • [IMPROVEMENT] Dogstatsd: Add support for SO_RCVBUF socket option. See #3429, thanks @jnewland
  • [IMPROVEMENT] JMXFetch: Touch JMXFetch launch file on boot-up. See jmxfetch-143
  • [IMPROVEMENT] Kubernetes: Add option to import node labels as host tags. See #3445
  • [IMPROVEMENT] Kubernetes: Automatically collect events from one agent only, with leader election among agents. See #3476
  • [IMPROVEMENT] Kubernetes: Tag pods with kube_stateful_set tag. See #3443, thanks @erkolson
  • [IMPROVEMENT] Packaging: Allow disabling datadog-agent service via sysconfig/default file. See #3301, #3483 and #3489, thanks @generica
  • [IMPROVEMENT] Packaging: On CentOS, add systemd service definition and add LSB dependency in sysvinit script. See #3477
  • [IMPROVEMENT] Prometheus: Allow metric type overrides in PrometheusCheck class. See #3463
  • [IMPROVEMENT] Windows packaging: Allow configuration of proxy on the installer CLI. See #3470 and dd-agent-omnibus-192
  • [IMPROVEMENT] Windows system checks: Use win32pdh instead of WMI. See #3294
  • [BUGFIX] Autodiscovery: Fix instance-level tags when multi instances are configured. See #3478
  • [BUGFIX] Autodiscovery: Re-pipe configurations after JMXFetch restart. See #3415
  • [BUGFIX] Collector: Only add service check fields if non-null. See #3487
  • [BUGFIX] Core: Make configcheck command exit with meaningful exit code. See #3481
  • [BUGFIX] Core: Make test on stored PID more accurate on Windows. See #3468
  • [BUGFIX] Core: Remove occurences where the full api key could be logged. See #3466
  • [BUGFIX] Forwarder: Make payload rejection log entry more explicit. See #3486
  • [BUGFIX] JMXFetch: Increase maximum length of instance configs pulled from Auto-Discovery pipe. See jmxfetch-147
  • [BUGFIX] JMXFetch: Fix Agent crash when JMXFetch pipe is broken. See #3461
  • [BUGFIX] Packaging: On CentOS, fix duplicate and unwanted service restarts at the end of installs. See dd-agent-omnibus-194
  • [BUGFIX] WMI: Convert UINT32 WMI type to positive integer instead of negative integer. See #3467
  • [SANITY] Autodiscovery: Avoid calling kubeutil if it's not initialized. See #3479
  • [SANITY] Collector: Downgrade docker warning message to debug on dockerless hosts. See #3485
  • [SANITY] Collector: Remove deprecated run_check method on Windows. See #3448
  • [SANITY] Dogstatsd: Parse distribution metrics and ignore them. See #3458
  • [SANITY] Gohai: Swallow python error for python version. See gohai-51
  • [SANITY] Packaging: Remove unneded chown on /usr/bin/dd-agent. See dd-agent-omnibus-193

5.16.0 / 07-26-2017

Linux, Docker and Source Install



Please refer to the Integrations Core Checks 5.16.0 tag for the list of changes on the core checks. Please refer to the Process Agent for more information on this new binary.

5.15.0 / 07-18-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.15.0 tag for the list of changes on the core checks. Please refer to the Trace Agent 5.15.0 tag for the list of changes on the Trace Agent.

  • [FEATURE] [kube] Allow to run agent on the node #3401
  • [FEATURE] [core] Adds fraction of file handles in use over system limit #3395
  • [FEATURE] [PrometheusCheck] add ignore_metrics array #3394
  • [FEATURE] pkg: Add process-agent to supervisor and flares #3409
  • [FEATURE] Get real host aliases for Cloud Foundry #3406
  • [FEATURE] [docker] add docker and orchestrator (ECS/Mesos/Nomad/Kubernetes/Swarm) version host tags. See #3383, #3386 and #3403
  • [FEATURE] [Windows] add additional calculator types #3378
  • [FEATURE] Mesos: extract marathon_app, chronos_job and chronos_job_owner tags. See #3375, #3414 and #3417
  • [FEATURE] [kubernetes] add kubernetes container name to metric tags #3368
  • [FEATURE] Add support for user/password authentication for the etcd autodiscovery store #3357
  • [FEATURE] Add support for %%pid%% template variables #3404, thanks @sophaskins and @jnewland
  • [IMPROVEMENT] constant for container name label #3405, thanks @sophaskins
  • [IMPROVEMENT] Add container name tag in auto discovery #3421, thanks @subuk
  • [IMPROVEMENT] Disable kube_service tagging on four 403 errors from apiserver #3413
  • [IMPROVEMENT] Update Mac restart command #3390
  • [IMPROVEMENT] Try both 4001 and 2379 etcd ports in SD if not configured #3351
  • [SANITY] Moves common wait_for_async network function out to the Agent #3408
  • [SANITY] [util] Move Timer to its own utils.timer module #3391
  • [SANITY] [checks] Deprecate AgentCheck.set #3385
  • [SANITY] Remove - no more use to us. #3364

5.14.1 / 06-15-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.14.1 tag for the list of changes on the core checks.

  • [IMPROVEMENT] Kubernetes: Reduce apiserver request count for pod_service_mapper. See #3387 and integrations-core-476
  • [IMPROVEMENT] CentOS: Make sure we start the last and stop first. See #3376
  • [IMPROVEMENT] Docker: Make TLS settings more intuitive. See #3373
  • [BUGFIX] Autodiscovery: Fix multi-instance autodiscovery. See #3341
  • [BUGFIX] Network checks: Fix typo in the network exception log. See #3374

5.14.0 / 06-05-2017

Linux, Windows, Docker and Source Install


Dependency changes


Please refer to the Integrations Core Checks 5.14.0 tag for the list of changes on the core checks. Please refer to the Trace Agent 5.14.0 tag for the list of changes on the Trace Agent.

  • [IMPROVEMENT] Autodiscovery metrics have new tags when available:
    • Docker image name and tag (#3350)
    • ECS task name (#3349 and #3358)
    • Nomad task, group and job names (#3289)
    • Kubernetes service and creator (deployment/daemon_set/replica_set/job). See #3300 and #3348
  • [IMPROVEMENT] New PrometheusCheck base class to enable simpler development of integrations pulling Prometheus format metrics. See #3317, #3344 and #3359, thanks @aerostitch
  • [IMPROVEMENT] Autodiscovery: allow multiple instances per check. See #3311
  • [IMPROVEMENT] AWS: bump boto version to 2.46.1. See #3307 and omnibus-software-137
  • [IMPROVEMENT] JMXFetch: Add support for min_collection_interval. See jmxfetch-135 and jmxfetch-140
  • [IMPROVEMENT] JMXFetch packaging: build JMXFetch with omnibus and store the jar on S3 bucket. See dd-agent-omnibus-100 and #3220
  • [IMPROVEMENT] Packaging: remove dependency on pycrypto. See dd-agent-omnibus-175
  • [IMPROVEMENT] Packaging: bump libxml2 version to 2.9.4. See omnibus-software-142
  • [IMPROVEMENT] Utils, Proxy: add a utility function to deal with proxy configs and skipping. See #3331 and #3356
  • [BUGFIX] Alpine: make supervisor socket consistent with default. See #3335, thanks @jonchiu
  • [BUGFIX] Collector: fix misattribution of check instance warning. See #3339
  • [BUGFIX] Core: respect ssl_skip_validation when checking API key. See #3346, thanks @jeffwidman
  • [BUGFIX] GCE: Exclude bosh_settings from tags, as it frequently includes passwords. See #3320
  • [BUGFIX] Packaging: remove agent/checks from PYTHONPATH. See #3315
  • [BUGFIX] Packaging: Fix == bashism in package scripts. See dd-agent-omnibus-174
  • [BUGFIX] Packaging: Fix dependency on windows DLLs on Windows 2008R2 by duplicating them into the embedded path. See dd-agent-omnibus-169
  • [SANITY] Dogstatsd: Log hostname on initialization. See #3332
  • [SANITY] JMXFetch: Add warnings of breaking change coming with agent 6. See #3318
  • [SANITY] Packaging: Remove checks deps in requirements.txt. See #3298

5.13.2 / 05-11-2017

Linux, Windows, Docker and Source Install



Please refer to the Integrations Core Checks 5.13.2 tag for the list of changes on the Integrations Core checks repository.

  • [IMPROVEMENT] Flare: add auto_conf templates. See #3330
  • [BUGFIX] Dockerutil: use DockerUtil.image_name_extractor for all code paths. See #3334
  • [BUGFIX] Dockerutil: fix exception when image name is None. See #3329
  • [BUFGIX] Dockerutil: support systemd slices in cgroup paths. See #3328
  • [BUGFIX] Dockerutil: fix image name when using sha256 for specs. See #3326
  • [BUGFIX] Packaging: fix os detection for "exotic" distributions. See dd-agent-omnibus-171
  • [SANITY] Packaging: Remove reliance on pedump fork. See dd-agent-omnibus-167

5.13.1 / 05-02-2017

Docker Alpine and Source Install Only



  • [BUGFIX] Redis: Fixes a bug where the redis python library would not install using the source install. See integrations-core-389

5.13.0 / 04-24-2017

Linux, Windows, Docker and Source Install


On this release the APM tracing agent is now enabled by default on supported platforms (linux). Should you wish to disable it please set:

apm_enabled: no

in the relevant section in datadog.conf and restart the agent.


Please refer to the Integrations Core Checks 5.13.0 tag for the list of changes on the Integrations Core checks repository. Please refer to the Trace Agent 5.13.0 tag for the list of changes on the Trace Agent.

  • [FEATURE] Docker, Service Discovery: Add ability to get docker IP address from Rancher labels for Rancher 1.3. See #3265
  • [FEATURE] Docker: improve network tagging. See #3272
  • [IMPROVEMENT] Kubeutil: catch errors from kubeutil init. See #3310
  • [BUGFIX] JMXFetch: fix race condition preventing SD initialization. See jmxfetch-135
  • [BUGFIX] Kubernetes_state: refactor gauge submission, fix container.restarts. See #3297
  • [BUGFIX] Dockerutil: fix net metrics in host mode. See #3313 and #3314
  • [SANITY] Windows: Remove system network check (metrics already reported by integrations-core). See #3096
  • [SANITY] Kafka: adding Kafka to list of JMX checks (NOP). See #3304
  • [REVERT] Core: SpooledTemporaryFile for subprocess output. See #3291

5.12.3 / 03-31-2017

Linux, Docker and Source Install



Please refer to the Trace Agent 5.12.3 tag for the list of changes on the Trace Agent.

  • [BUGFIX] Docker: Fix CGroup parsing. See #3293

5.12.2 / 03-29-2017

Linux, Windows, Docker and Source Install



  • [BUGFIX] Cloud Metadata: Use Metadata-Flavor to ensure that tags are properly collected on GCE. See #3288
  • [BUGFIX] Forwarder: Fix error on removal of transactions. See #3287
  • [BUGFIX] Docker: Fix cgroup parsing on RancherOS. See #3274

5.12.1 / 03-23-2017

Linux, Windows, Docker and Source Install



  • [BUGFIX] Source Install: Fixes a bug in the source install where the integrations wouldn't be properly installed. See #3281
  • [BUGFIX] Emitter: Force source type to system for legacy payload. See #3280
  • [BUGFIX] MySQL: Allow for configurable collection of replica statuses integrations-core-288
  • [BUGFIX] MySQL: Slaves_connected should be a gauge. integrations-core-291

5.12.0 / 03-22-2017

Linux, Windows, Docker and Source Install


Integrations SDK

In all agents, we are now able to release independent check packages, so that we can update checks independently and also release 3rd party packages from the community. To aid in this, we now have three repos: the agent repo, the core integrations repo, and the extras integrations repo. Core integrations are officially supported by Datadog, extras are supported by the community.

Windows packaging changes

The Windows agent is now always installed as a "per-machine"; previous versions installed as "per-user" by default. Upgrades from "per-user" to "per-machine" are performed using the executable (.exe) distribution. Please refer to the dedicated page on the wiki for detailed upgrade instructions.

The Windows agent is now compatible with 64-bit platforms only.

The Windows agent is now built with Omnibus. It now ships a full python interpreter, easing the install of additional python libraries for custom checks.

The Windows agent is now digitally signed.

New Integrations

Updated integrations

  • ActiveMQ
  • Consul
  • Couchbase
  • Disk
  • DNS Check
  • Docker
  • Docker Daemon
  • Elasticsearch
  • Etcd
  • JMXFetch
  • Http Check
  • HAProxy
  • IIS
  • Kubernetes
  • Marathon
  • Mesos
  • Mongo
  • MySQL
  • Network
  • OpenStack
  • PowerDNS
  • PGBouncer
  • Postgres
  • Process
  • Redis
  • Spark
  • SQLServer
  • Riak
  • TCP Check
  • TokuMX
  • VSphere
  • Windows
  • Yarn
  • Zookeeper

Dependency Upgrades


  • [FEATURE] Core: Adds a local DNS cache. See #3197
  • [FEATURE] JMXFetch: Add histogram metric type. See jmxfetch-115
  • [FEATURE] JMXFetch: Add list_jvms command to list available JVMs when using the Attach API. See jmxfetch-100, jmxfetch-112 (Thanks @cslee00)
  • [FEATURE] JMXFetch: Add tag blacklisting. See jmxfetch-116
  • [FEATURE] JMXFetch: Add user tags definition for MBeans. See jmxfetch-117.
  • [FEATURE] JMXFetch: Enable service discovery via a named pipe. See jmxfetch-113
  • [FEATURE] JMXFetch: Support attribute types. See jmxfetch-111, jmxfetch-128 (Thanks @brothhaar)
  • [FEATURE] JMXFetch: Support user tag value substitution by attribute name. See jmxfetch-117.
  • [FEATURE] Service Discovery, JMX: Allows JMX Checks to use Service Discovery. See #3010, #3150
  • [FEATURE] Service Discovery: Add a Zookeeper backend. See #3038
  • [FEATURE] Service Discovery: Use a template cache to reduce calls to template store. See #3060
  • [FEATURE] Service Discovery: Add namespace, daemon_set and replica_set tags for kubernetes. See #3176
  • [FEATURE] Windows: Adds new service and packaging. See #2417, #3244
  • [IMPROVEMENT] Core: Allows developer mode to be used without profiling. See #2898 (Thanks @cberry777)
  • [IMPROVEMENT] Core: Better EC2 detection. See #3174
  • [IMPROVEMENT] Core: Splits the payload into a legacy payload and a metrics payload. See #3180, #3240
  • [IMPROVEMENT] Forwarder: Validate API Key against any endpoint. See #3256
  • [IMPROVEMENT] JMXFetch: Print exception messages on Attach API connection failures. See jmxfetch-122 (Thanks @aoking)
  • [IMPROVEMENT] Packaging: Don't require pgp key to be added on every install. dd-agent-omnibus-143
  • [IMPROVEMENT] Windows: Better pagefile memory on Windows. See #3072
  • [BUGFIX] Aggregator: Parse tags even if sample rate is bad. See #3073
  • [BUGFIX] Core: Fix config parsing when variable is missing. See #3041
  • [BUGFIX] Core: Exclude sensitive data from Google Cloud metadata. See #3076
  • [BUGFIX] Core: Stops agent from dying when no proc table is returned. See #3043 (Thanks @ahamilton55)
  • [BUGFIX] Flare: Fix config file parsing. See #3040, #3059
  • [BUGFIX] JMXFetch: Allow specifying no alias on detailed attribute. See jmxfetch-133
  • [BUGFIX] JMXFetch: Fix connectivity loss when multiple instances are assigned to a same JVM. See jmxfetch-124
  • [BUGFIX] JMXFetch: Parse string-defined ports to integers in user configurations. See jmxfetch-121
  • [BUGFIX] JMXFetch: Support java.util.Map attribute types. See jmxfetch-130
  • [BUGFIX] JMXFetch: Support list-defined user tags at instance level. See jmxfetch-132
  • [BUGFIX] Windows GUI: Better detection and termination of old processes on startup. See #3125
  • [DEPRECATE] Core: hard-deprecate start/stop/restart/status commands. See #3004

Integration Changes

  • [FEATURE] DNS Check: Change to a network check. See #2924
  • [FEATURE] Docker Daemon: Add a number of new metrics for container and volume counts. See #2740, #3077. (Thanks @parkr)
  • [FEATURE] Docker Daemon: Tag container metrics with swarm if available. See #3182, #3243
  • [FEATURE] Http Check: Mark the Service Check as down if result matches content match option. See #3069
  • [FEATURE] HAProxy: Allow the use of Unix Socket URLs. See #3005 (Thanks @sj26)
  • [FEATURE] HAProxy: Add support for haproxy.backend.uptime. See #2639
  • [FEATURE] Kubernetes: Allow kublet host to be specified from environment. See #3051 (Thanks @stonith)
  • [FEATURE] Kubernetes: Add image_name and image_tag to container metrics. See #2990 (Thanks @tarvip)
  • [FEATURE] Kubernetes State: Add State metrics for pods and containers. See #3070. (Thanks @therc)
  • [FEATURE] Marathon: Add token auth. See #3208, #3037. (Thanks @jkorzekwa-atlassian)
  • [FEATURE] Mongo: Add metric with number of databases. See #3124. (Thanks @mpidlisnyi)
  • [FEATURE] Network: Add InCsumErrors to UDP metrics. See #3117. (Thanks @diver-in-sky)
  • [FEATURE] Network: Adds an option to split apart connection states. See #3158, #2856. (Thanks @joewilliams)
  • [FEATURE] PowerDNS: Add support for v4. See #3166, #3066. (Thanks @jimmystewpot)
  • [FEATURE] Redis: Add some additional metrics. See #2749
  • [FEATURE] SQLServer: Allow connection through pyodbc as well as adodbapi. See integrations-core-259, integrations-core-264, omnibus-software-129, dd-agent-omnibus-154
  • [FEATURE] Riak: Add support for Riak Security. See: #2389, #3168. (Thanks @hamano)
  • [FEATURE] Riak: Add support for Riak 2.0 LTS. See: #3189, #2762. (Thanks @jcapricebasho)
  • [FEATURE] VSphere: Improve tagging, v6 compatibility, infrastructure reporting and performance. See #3055
  • [IMPROVEMENT] Consul: Add configuration parameter that overrides MAX_SERVICES. See integrations-core-140
  • [IMPROVEMENT] Couchbase: Add user specified tags to service checks. See #3079. (Thanks @arzarif)
  • [IMPROVEMENT] DNS Check: Support monitoring performance of NXDOMAIN queries. See #2849. (Thanks @jnewland)
  • [IMPROVEMENT] Docker: Improved Hostname logic for containers. See #3116
  • [IMPROVEMENT] Docker: Better detection and handling of incorrect PIDs. See #3218, integrations-core-237
  • [IMPROVEMENT] Docker Daemon: Make Docker Healthcheck a Service Check. See #2859
  • [IMPROVEMENT] Etcd: Report errors connecting to etcd endpoint. See #3007 (Thanks @pbitty)
  • [IMPROVEMENT] HTTP Check: Adds optiont to disable default http headers. See integrations-core-182. (Thanks @eredjar)
  • [IMPROVEMENT] HTTP Check: Remove noisy debug logging. See integrations-core-267
  • [IMPROVEMENT] Kubernetes: handle multiple namespaces. See #3028
  • [IMPROVEMENT] Kubernetes: Support api server auth with a cert. See #3145
  • [IMPROVEMENT] Kubernetes: Allow configurable custom certs. See #3160
  • [IMPROVEMENT] Kubernetes: Add support for kublet auth when the read-only port is disabled. See #3142, integrations-core-242
  • [IMPROVEMENT] Kubernetes: Updates the path to grab certs. See #3210. (Thanks @dturn)
  • [IMPROVEMENT] Marathon: Add option to disable SSL validation and support groups. See #3140. (Thanks @Carles-Figuerola)
  • [IMPROVEMENT] Mesos: Adds mesos.registrar.log.recovered to metrics. integrations-core-260
  • [IMPROVEMENT] MySQL: Add another format for innodb writes innodb. See #3148
  • [IMPROVEMENT] Network: Use a global counter for thread pool size. See #3095, #3080
  • [IMPROVEMENT] OpenStack: Support regex for Network ID exclusion. See integrations-core-181
  • [IMPROVEMENT] PHP-FPM: Include Custom Tags in Service Check. See #3109
  • [IMPROVEMENT] PHP-FPM: Add http_host tag to metrics. See #3165, #3074. (Thanks @toksvaeth)
  • [IMPROVEMENT] Postgres: Makes db size collection optional. See #3035 (Thanks @jstotzs)
  • [IMPROVEMENT] Spark: No events on job status change. See #3194. This is a potentially breaking change, but it was flooding event streams with what, for most people, was useless information.
  • [IMPROVEMENT] TCP Check: add instance tags to respose_time metric. See #3118, #3206. (Thanks @mberner-ch)
  • [IMPROVEMENT] TokuMX: Skip dbs that aren't accessible. See #3093. (Thanks @pbrisbin)
  • [IMPROVEMENT] Yarn: Add configurable application tags. See #3041, integrations-core-261
  • [IMPROVEMENT] Zookeeper: Include user specified tags. See #3078. (Thanks @arzarif)
  • [IMPROVEMENT] Trace Agent: Watchdog restarts trace agent if it consumes too much memory or opens too many network connections. trace-trace-agent-182
  • [IMPROVEMENT] Trace Agent: Reduce sampler log verbosity and expose internal state with Go expvar. datadog-trace-agent-248
  • [IMPROVEMENT] Trace Agent: Adaptive sampling. datadog-trace-agent-228
  • [IMPROVEMENT] Trace Agent: Weight spans using their sample rate. datadog-trace-agent-226
  • [IMPROVEMENT] Trace Agent: Stream decoding. datadog-trace-agent-227
  • [IMPROVEMENT] Trace Agent: Quantize SQL AS aliases. trace-trace-agent-225
  • [IMPROVEMENT] Trace Agent: Updated the output of trace-agent info. datadog-trace-agent-219
  • [BUGFIX] Disk: Tag metrics with filesystem tag key instead of filesystem itself. See integrations-core-252
  • [BUGFIX] Docker: Fix whitelist pattern matching. See #3048
  • [BUGFIX] Docker: Fix image tag extraction. See #3172
  • [BUGFIX] Docker Daemon: Report as many cgroup metrics as possible. See #3134
  • [BUGFIX] Elasticsearch: Fix autoconf to use the correct port. See #3129
  • [BUGFIX] Http Check: Fix content match for non ascii characters. See #3100
  • [BUGFIX] IIS: Fix is_2008 option. See #3039
  • [BUGFIX] Kubernetes: Only use annotations for service discovery once per pod. See #2901 (Thanks @mikekap)
  • [BUGFIX] Kubernetes: Fix tags param in example config file. See #3044
  • [BUGFIX] Kubernetes: Remove potentially sensitive information from logs. See integrations-core-254
  • [BUGFIX] Marathon: Catch nodata found errors and report them properly. See #3127
  • [BUGFIX] MySQL: Use information_schema in versions below 5.6.0. See #3196
  • [BUGFIX] MySQL: Fix version comparison operator. See integrations-core-231
  • [BUGFIX] OpenStack: Fallback to v2 endpoint if v2.1 fails. See #3154
  • [BUGFIX] PGBouncer: Disable PGBouncer on Windows, as it doesn't work. See #3252
  • [BUGFIX] Process: No longer errors when process is down. See #3052 and #3049 (Thanks @Radstake)
  • [BUGFIX] Process: Send all configured tags with process check. See #2976 (Thanks @jdavisp3)
  • [BUGFIX] Spark: Properly report job IDs. See #3111
  • [BUGFIX] Spark: Fix event source name. See #3193
  • [BUGFIX] SQLServer: Stops passwords from leaking into logs. See #3053
  • [BUGFIX] SQLServer: Collect metric list if SQLServer is not up during check init. See #3067
  • [BUGFIX] Trace Agent: Now shells out to dd-agent to retrieve hostname. datadog-trace-agent-242
  • [BUGFIX] Trace Agent: Normalize environment strings in spans and in the configuration datadog-trace-agent-233
  • [BUGFIX] Trace Agent: Fixed race issue between concentrator and sampler. datadog-trace-agent-229
  • [DEPRECATE] ActiveMQ: Remove old JMX config file. See #3141

5.11.3 / 02-27-2017

Mac OS only


There was an issue with Gohai on Mac OS >= 10.11 leading to high CPU utilization. As a short term solution, Gohai has been disabled on Mac OS.


  • [OTHER] Do not run Gohai on Mac OS. See #3224.

5.11.2 / 02-13-2017

Linux and Docker only


There was an issue releasing 5.11.1 by which the package shipped still contained the bugs reported in 5.11.0 plaguing the SNMP check. Only affected debian package and the docker release.


  • [RELEASE] Releasing actual bugfix expected with 5.11.1, with version bump to 5.11.2.

5.11.1 / 02-13-2017

Linux and Docker only


There was a regression in 5.11.0 due to an inadverted subdependency upgrade, breaking the SNMP check.


5.11.0 / 02-08-2017

Linux and Docker only


This release bundles the datadog-trace-agent as part of Linux and Docker installs. The dd-agent codebase contains a minimal set of changes to ensure the new trace-agent process will plug neatly into existing agent installs. Beta customers of Datadog's tracing tool will have received an email detailing the steps to transition to dd-agent 5.11.0 from the deprecated dd-trace-agent package.

For more information:


  • [IMPROVEMENT] Supervisor: Add trace-agent as an optional process managed by supervisor
  • [IMPROVEMENT] Flare: Add trace-agent log file to flare output
  • [IMPROVEMENT] Core: append the output of trace-agent -info to service datadog-agent info command

5.10.1 / 11-21-2016

Linux, Windows, Docker and Source Install


Updated Integrations

  • RiakCS
  • Mongo


  • [IMPROVEMENT] Core/Forwarder: stop flushing after 10s. See #3018.

  • [IMPROVEMENT] Core: isolate system checks. See #3011.

  • [IMPROVEMENT] RiakCS: support Riak CS 2.1+ stats format. See #2920. (Thanks @millerdev)

  • [IMPROVEMENT] Status: Silence requests exception. See #3023.

  • [IMPROVEMENT] SpooledTemporaryFile for subprocess output. See #3002.

  • [BUGFIX] Core: fix unintended subprocess_output empty output errors. See #3024.

  • [BUGFIX] Core/Multiple Checks: Only set psutil.PROCFS_PATH once in the collector. See #3013.

  • [BUGFIX] Core: use proxy for API key status check in info page. See #3012. (Thanks @2rs2ts)

  • [BUGFIX] Mongo: use db.current_op instead of manually querying. See #3016 (Thanks @ebroder)

  • [BUGFIX] Mongo: use currentOp for monodb 3.1+. See #3015 (Thanks @lattwood)

  • [DEPRECATE] Process: procfs_path is now deprecated, should be set in datadog.conf. See #3013.

5.10.0 / 11-09-2016

Linux, Windows, Docker and Source Install


Kubernetes: metrics from the Kube State API

The new kubernetes_state check pulls metrics from the Kubernetes State API.

See #2890, #2984 and omnibus-software-85 for more details.

RPM package for SUSE

The Agent is now available as an RPM package for SUSE. Please follow the instructions on our Agent installation page to use it.

For more details, see related PRs #2900, #2937, omnibus-software-72, dd-agent-omnibus-90

Linux packages: change of ownership of /opt/datadog-agent directory

The Linux packages (deb and rpm) now make the dd-agent user and group own the /opt/datadog-agent folder and its contents when they're installed. This shouldn't cause any change on regular installations of the Agent, but please take this change into account if you customize the ownership/permissions of the /opt/datadog-agent directory or its contents.

For more details, see related PR: dd-agent-omnibus-97

Python dependencies upgrade

We have updated the following python dependencies shipped with the Agent. If you happen to run any custom checks that rely on the former versions please make necessary amends:

Updated integrations

  • Apache
  • Ceph
  • Consul
  • Docker
  • Elasticsearch
  • FluentD
  • Haproxy
  • Http check
  • Kafka
  • Kubernetes
  • Marathon
  • Mesos
  • Mongo
  • MySQL
  • Openstack
  • PGBouncer
  • Postgres
  • Process
  • RabbitMQ
  • Redis
  • Spark
  • SSH check


  • [FEATURE] Core: Add support for sum histogram_aggregates type. See #2780 (Thanks @ross)
  • [FEATURE] Core: Display api key status on the info page. See #2947
  • [FEATURE] Consul: Add net latency metrics. See #2757 (Thanks @ross)
  • [FEATURE] Haproxy: Allow per-instance custom tags. See #2649, #2951 (Thanks @theojulienne)
  • [FEATURE] Kubernetes: Add CPU and memory capacity reporting. See #2935 (Thanks @markine)
  • [FEATURE] Mongo: Add metrics for fsyncLocked-ness, replset votes and vote fraction. See #2975 (Thanks @ebroder)
  • [FEATURE] Process: Support pid file. See #2607 (Thanks @n0ts)
  • [FEATURE] Spark: Support clusters provisioned using Apache Mesos. See #2797 (Thanks @zachradtka)
  • [IMPROVEMENT] Ceph: Handle HEALTH_WARNING status. See #2929 (Thanks @ccocsas)
  • [IMPROVEMENT] Consul: Remove warning when no whitelist is defined. See #2955
  • [IMPROVEMENT] Consul: Support multiple instances. See #2948
  • [IMPROVEMENT] Core: forwarder: Do not requeue all responses. See #2840
  • [IMPROVEMENT] Core: collector: Catch exceptions raised by system checks. See #2973
  • [IMPROVEMENT] Docker: Set alert_type to error for oom and kill events. See #2866
  • [IMPROVEMENT] Docker: Set exec events' priority to low. See #2960
  • [IMPROVEMENT] Docker/Kubernetes: Filter containers consistently. See #2836
  • [IMPROVEMENT] Elasticsearch: Add support for ES 5.x. See #2989 (Thanks @ptinsley and @alexef)
  • [IMPROVEMENT] FluentD: Add collection timeout. See #2978 (Thanks @mattrobenolt)
  • [IMPROVEMENT] Kafka: Document specifying multiple Kafka / ZK nodes. See #2956 (Thanks @jeffwidman)
  • [IMPROVEMENT] Kafka: Upgrade kazoo dep to 2.2.1. See #2964, omnibus-software-84
  • [IMPROVEMENT] Kubernetes: Use the env var KUBERNETES_SERVICE_HOST if present. See #2921 (Thanks @arohner)
  • [IMPROVEMENT] Kubernetes: Add namespace tag to kubernetes.pods.running. See #2931
  • [IMPROVEMENT] Kubernetes: Add some tags to events. See #2926
  • [IMPROVEMENT] Marathon: Count running/pending deployments. See #2788 (Thanks @bradhe)
  • [IMPROVEMENT] Packaging: Upgrade psutil to 4.4.1. See #2957, omnibus-software-82
  • [IMPROVEMENT] Packaging: Upgrade version of datadogpy shipped with Linux packages. See omnibus-software-80
  • [IMPROVEMENT] PGBouncer: Support SSL connections to PGBouncer. See #2885, #2967 (Thanks @dblackdblack)
  • [IMPROVEMENT] Postgres: Allow UNIX socket path as host config value. See #2734, #2968 (Thanks @DylanFrese)
  • [IMPROVEMENT] RabbitMQ: Add option to skip SSL verification. See #2991 (Thanks @iancward)
  • [IMPROVEMENT] Redis: Fix tags on can_connect service check. See #2803 (Thanks @gphat)
  • [IMPROVEMENT] Redis: Enable monitoring of SSL-ified redis servers. See #2674 (Thanks @dblackdblack)
  • [IMPROVEMENT] SSH Check: Support ECDSA keys (<= 256bits). See #2965 (Thanks @n0ts)
  • [BUGFIX] Apache: Add timeout on connection to Apache status URL See #2928
  • [BUGFIX] Core: Avoid CPU spikes caused by subprocess calls. See #2950
  • [BUGFIX] Core: Catch EOFError when loading info page. See #2972
  • [BUGFIX] Core: Display expected service check count. See #2938
  • [BUGFIX] Core: Handle timezone encodings on Windows. See #2942
  • [BUGFIX] Core: Ignore all non-unicode-decodable characters in collector's emitter. See #2941
  • [BUGFIX] Core: In dev mode, fallback to /tmp to dump pstats. See #2934
  • [BUGFIX] Core: On the Windows UI, set color to white for background of info page. See #2970
  • [BUGFIX] Core: On Windows, retry on process restart failures. See #2897
  • [BUGFIX] Kubernetes: Fix kubelet health check request. See #2887 (Thanks @yannrouillard)
  • [BUGFIX] Kubernetes: Fix parsing of units in Kubernetes resource quantities. See #2908 (Thanks @therc)
  • [BUGFIX] Http Check: Fix SSL certificate parsing issue related to strptime. See #2915 (Thanks @ghTravis)
  • [BUGFIX] Mesos Slave: Connect to master with same protocol as slave. See #2910 (Thanks @mcclurmc)
  • [BUGFIX] Service Discovery: Fix single check reload on auto_conf templates and removed containers. See #3006
  • [BUGFIX] Mongo: Fix case when check would send password in clear. See #2940
  • [BUGFIX] Mongo: Fix hostname detection when username/password are part of the uri. See #2894 (Thanks @ibwhite)
  • [BUGFIX] MySQL: Improve innodb engine status check on RDS Aurora. See #2971 (Thanks @scottgeary)
  • [BUGFIX] MySQL: Cast password to string to avoid issues with number-only passwords. See #2983 (Thanks @michalyao)
  • [BUGFIX] Openstack: Add missing SSL verify param to 2 requests calls. See #2877 (Thanks @jcejohnson)
  • [BUGFIX] Windows packaging: Terminate agent-manager.exe on uninstall/upgrade. See #2982

5.9.1 / 09-30-2016

Linux, Docker and Source Install


Build/Packaging Issue in 5.9.0

A packaging issue led to issues with the Spark integration when running Agent 5.9.0. This release fixes the issue.

Updated integrations

  • Kubernetes
  • Spark


  • [BUGFIX] Dogstatsd: fix ipv6 fallback to ipv4. See #2879
  • [BUGFIX] Kubernetes: Fix incorrect time conversion for Kubernetes events. See #2873 (Thanks @yannrouillard)
  • [BUGFIX] Packaging: Build with the corresponding dd-agent-omnibus branch ( 5.9.x )
  • [BUGFIX] Packaging: include latest repo keys. See above.
  • [BUGFIX] Spark: fix missing libraries issue.

5.9.0 / 09-27-2016

Linux, Windows, Docker and Source Install


Kafka-Python library update

Because we have bumped the library version for the kafka-python package, if you happen to run any custom checks that rely on the former version please make necessary amends. We apologize for any inconvenience. See #2709 for reference.

Updated integrations

  • Apache
  • Cassandra
  • Ceph
  • Disk Check
  • DNS
  • Docker
  • Elasticsearch
  • Gearman
  • HTTP Check
  • IIS
  • JMXFetch
  • Kafka Consumer
  • Kubernetes
  • Linux Check
  • Marathon
  • Mesos
  • Mongo
  • MySQL
  • Network Check
  • OpenStack
  • Php5_fpm
  • Postgres
  • Process
  • RabbitMQ
  • Redis
  • Spark
  • SSH
  • vSphere
  • Windows

Deprecated integrations

  • Jenkins - Replaced check with plug-in. Details below.


  • [ENHANCEMENT] Use JMXFetch 0.12.0 ( Changelog )

  • [SECURITY] The deb and rpm packages now bundle OpenSSL 1.0.2i. For more details, see the security advisory, omnibus-software-76

  • [FEATURE] Apache: add connection status metrics. See #2622 (Thanks @gzysk8)

  • [FEATURE] Core: dogstatsd: Added support for IP agnostic connections. See #2619

  • [FEATURE] Curl: support for HTTP2. See omnibus-software-65

  • [FEATURE] Dns_check: Allow querying a custom record type. See #2653

  • [FEATURE] Docker_daemon: resolve hierarchy by looking at procfs /cgroup. See #2537

  • [FEATURE] Gearman: Add per-task gearman metrics. See #2672 (Thanks @nyanshak)

  • [FEATURE] Http_check: Add allow_redirects parameter. See #2586 (Thanks @n0ts)

  • [FEATURE] Http_check: Add ability to use post method. See #2544

  • [FEATURE] Http_check: SNI support for cert_expiration. See #2521

  • [FEATURE] Kubernetes: Improve service discovery to only reload checks that need it. See #2551

  • [FEATURE] Kubernetes: Ingest k8s events + limits and requests metrics. See #2551

  • [FEATURE] Linux: make /proc configurable. See #2482

  • [FEATURE] Mongo: Apply yaml-configured tags to service checks. See #2575 (Thanks @avaughan89)

  • [FEATURE] Mysql: Allow connection_timeout to be set for MySQL checks. See #2729 (Thanks @scottgeary)

  • [FEATURE] Process: Option to search processes by PID. See #2119 (Thanks @handigarde)

  • [FEATURE] Spark: spark standalone support. See #2752 (Thanks @zachradtka), omnibus-software-73, dd-agent-omnibus

  • [IMPROVEMENT] Cassandra: add read and write latency ColumnFamily metrics. See #2483

  • [IMPROVEMENT] Ceph: improve health metrics. See #2852

  • [IMPROVEMENT] Ceph: update to support See #2805

  • [IMPROVEMENT] Core: avoid calls to service discovery from dogstatsd. See #2798

  • [IMPROVEMENT] Core: Add an HTTP timeout to many integrations. See #2673 (Thanks @gphat)

  • [IMPROVEMENT] Core: expiration should adjust to longer collection intervals (>300s). See #2541

  • [IMPROVEMENT] Core: remove events & resources api_key. See #2557

  • [IMPROVEMENT] Core: remove noisy logs. See #2715 (Thanks @ensonik)

  • [IMPROVEMENT] Disk: handle multilines df output. See #2733

  • [IMPROVEMENT] Docker: Improve service discovery to only reload checks that need it. See #2702

  • [IMPROVEMENT] Docker: Add CPU throttling stats to Docker metrics. See #2724 (Thanks @ejholmes)

  • [IMPROVEMENT] Dogstatsd: fix server address when non_local_traffic is passed. See #2691

  • [IMPROVEMENT] Elasticsearch: Add missing metrics. See #2758 (Thanks @mdelaney)

  • [IMPROVEMENT] Elasticsearch: Add elastic search cluster name to tags. See #2744, #2820 and #2826

  • [IMPROVEMENT] Flare: Close the tar file cleanly once upload is done. See #2621

  • [IMPROVEMENT] Gohai: Processes - simplify payload. See #2600

  • [IMPROVEMENT] Gohai: CPU - take into account environments with multiple physical processors. See gohai-32

  • [IMPROVEMENT] Http_check: log exceptions 🔊. See #2704

  • [IMPROVEMENT] Kafka_consumer: bumping kafka-python package version to 1.2.5. See #2709, omnibus-software-70

  • [IMPROVEMENT] Kubernetes: Add ability to specify service discovery via kubernetes annotations. See #2848 (Thanks @mikekap)

  • [IMPROVEMENT] Kubernetes: disable use_histogram. See #2542

  • [IMPROVEMENT] Kubernetes: Log URL in kubelet check failures and service checks. See #2735 (Thanks @therc)

  • [IMPROVEMENT] Mesos: make SSL check optional. See #2809

  • [IMPROVEMENT] Mongo: Adds metrics for collections. See #2739

  • [IMPROVEMENT] Mongo: Support X.509 authentication. See #2713

  • [IMPROVEMENT] Multiple integrations: Add an HTTP timeout to many integrations. See #2673 (Thanks @gphat)

  • [IMPROVEMENT] Network: implement check using psutil on Windows. See #2499

  • [IMPROVEMENT] Openstack: openstack should pick-up on proxy settings set on agent. See #2572

  • [IMPROVEMENT] Openstack: nova/neutron extend name search. See #2627

  • [IMPROVEMENT] Packaging: debian - only create dd-agent group if it doesn't already exist. See dd-agent-omnibus-89

  • [IMPROVEMENT] Packaging: linux - correctly silence configcheck. See dd-agent-omnibus-86

  • [IMPROVEMENT] Packaging: debian - remove configcheck output when upgrading. See dd-agent-omnibus-85

  • [IMPROVEMENT] Php5_fpm: Allow host header override. See #2779 (Thenks @allixsenos)

  • [IMPROVEMENT] Postgres: Allow to use psycopg2. See #2782

  • [IMPROVEMENT] Process: Add metrics for avg, max, and min run times for processes. See #2531 (Thenks @garnermccloud)

  • [IMPROVEMENT] Process: Add system.processes.mem.pct metric to process check. See #2801

  • [IMPROVEMENT] Python: Upgrade to 2.7.12. See omnibus-software-63

  • [IMPROVEMENT] RabbitMQ: Avoiding the extra double lookup for 'rabbitmq_api_url'. See #2543 (Thanks @ParthKolekar)

  • [IMPROVEMENT] RabbitMQ: new health check. See #2711

  • [IMPROVEMENT] SSH: Be more specific when logging ssh errors. See #2708

  • [IMPROVEMENT] Util: remove LaconicFilter. See #2605

  • [IMPROVEMENT] Vsphere: Add optional vm include parameter. See #2459

  • [BUGFIX] Core: forwarder drop payloads when no valid endpoints. See #2833

  • [BUGFIX] Core: emitter would crash agent with some unicode control characters in payload. See #2843

  • [BUGFIX] Core: service discovery fix fallback IP address extraction. See #2855

  • [BUGFIX] Core: service discovery dont pass missing config_store reference to dockerutil. See #2858

  • [BUGFIX] Cloud metadata: Fail on non-2xx responses. See #2844

  • [BUGFIX] Disk: timeout on disk usage. See #2714 and #2823

  • [BUGFIX] Docker_daemon: ECS introspection resilience. See #2745 and #2825

  • [BUGFIX] Docker_daemon: Custom cgroups are per instance only. See #2846

  • [BUGFIX] Docker_daemon: Fix pid retrieval that could affect k8s. See #2847

  • [BUGFIX] Dockerutil: cpu subsys might be mistaken for cpuacct. See #2829

  • [BUGFIX] Http_check: Bring back include_content option. See #2631

  • [BUGFIX] IIS: Fix metrics tagging when multiple sites are specified on instance. See #2677

  • [BUGFIX] JMXFetch: Do not scope MBeans queries on list_not_matching_attributes action. See jmxfetch-102

  • [BUGFIX] Marathon: Fix a small problem that prevented marathon full path from being properly built. See #2620

  • [BUGFIX] Mongo: Repairing of mongodb.can_connect check. See #2658 (Thanks @cryptspirit)

  • [BUGFIX] Mongo: Fixes a mistake in the mongo collections check. See #2783

  • [BUGFIX] MySQL: Check for null key_buffer_size before dividing by it. See #2784

  • [BUGFIX] Network Checks: Network check instance corrections. See #2736 (Thanks @tschellhorn)

  • [BUGFIX] Packaging [linux]: Kernel 4+ won't error out on status now. See #2640

  • [BUGFIX] Packaging [osx]: Support GNU-Friendly sed. See #2580 (Thanks @sha1sum)

  • [BUGFIX] Process: Use configured proc path when getting pagefault stats #2727 (Thanks @rdxiang)

  • [BUGFIX] Redisdb: Avoid exception when slowlog query returns empty command. See #2651

  • [BUGFIX] Win32_event_log: Fix small regression on msg_text selection. See #2591

  • [BUGFIX] Win32_event_log: reset filters between iterations. See #2796

  • [BUGFIX] Wmi: set provider architecture when necessary. See #2812

  • [DEPRECATED] Core: clean debug logs and kill Dogstream-based DdForwarder. See #2628

  • [DEPRECATED] Jenkins: Deprecated check. Please use plugin. More information can be found on the Jenkins Integration panel under the Configuration tab ( See #2688

  • [DEPRECATED] Nghttp2: remove python bindings. See omnibus-software-71

5.8.5 / 07-27-2016

Windows, Linux and Source Install


Future rotation of the APT signing key

In preparation for a future rotation of our package signing keys, the 5.8.5 DEB package will, on install, import a new trusted APT key:

pub   4096R/382E94DE 2016-06-29 [expires: 2022-06-28]
uid                  Datadog, Inc <package[@datadoghq][].com>

During the package install, the DEB package will output the following:

Prepare Datadog Agent keys rotation
  Add the new 'Datadog, Inc <package[@datadoghq][].com>' key to the list of APT trusted keys. ... OK

The signing key of the Agent hasn't changed yet but will be switched to this new key in a future release.

See dd-agent-omnibus-81


  • [IMPROVEMENT] Core: Upgrade embedded python to 2.7.12. See omnibus-software-63

  • [IMPROVEMENT] Elasticsearch: Take into account node name for cluster stats. See #2696

  • [IMPROVEMENT] RPM package: Add runtime dependency on the initscripts package. See dd-agent-omnibus-82

  • [IMPROVEMENT] RPM package: Don't delete dd-agent user and group on uninstall. See dd-agent-omnibus-84

  • [BUGFIX] Core: Use flag to signal config reload to avoid race conditions. See #2670

  • [BUGFIX] Core: Don't use Docker hostname if it's an EC2 one. See #2661

  • [BUGFIX] Docker: Fix disk metrics rounding issue. See #2626

  • [BUGFIX] Haproxy: Fix KeyError when an unknown status is found. See #2681

  • [BUGFIX] IIS: Remove warnings on 'Name' property. See #2633

  • [BUGFIX] MongoDB: Fix case where optimeDate is not available. See #2625

  • [BUGFIX] Service Discovery: Introduce _get_image_ident and fix a bug that needed it. See #2684

  • [BUGFIX] Windows Event log: Support unicode messages. See #2660

5.8.4 / 07-08-2016

Windows, Linux and Source Install



  • [IMPROVEMENT] Core: Upgrades supervisord to 3.3.0 See #2667, #2669

  • [BUGFIX] MySQL: Fixes MySQL slave detection. See #2610

  • [BUGFIX] MySQL: Fixes MySQL replication service check. See #2603

  • [BUGFIX] Core: Fixes a bug that caused the thread pool to leak threads. See #2666

5.8.3 / 07-05-2016

Windows, Linux and Source Install



  • [FEATURE] Flare: Adds configcheck output to flare command. See #2588

  • [IMPROVEMENT] Docker: Bump docker-py to 1.8.1 for network features support. See #2556

  • [BUGFIX] HAProxy: Add collate_status_tags_per_host flag See #2590

  • [BUGFIX] Core: A bug prevented ipv6 from being used in some circumstances. This fixes that, and defaults to ipv6 and falls back to ipv4. See #2592

  • [BUGFIX] Docker: Handles buggy responses from docker api better. See #2608

  • [BUGFIX] MongoDB: Only collect ReplicationInfo when it's possible to do so, rather than erroring out. See #2559

  • [BUGFIX] Postgres: Adds a hard limit on postgres.table.count metric, as this can be very expensive. See #2575

  • [BUGFIX] PowerDNS Recursor: The configuration file needed to be renamed from powerdns.conf to powerdns_recursor.conf. See #2538

  • [BUGFIX] Service Discovery: Improvements for testing, logging and service variable interpolation. See #2573

  • [BUGFIX] Service Discovery: Yse docker hostname rather than default route to query cadvisor and kublet. See #2609

  • [BUGFIX] Service Discovery: Use get_identifier instead of buggy image name extraction. See #2593

  • [BUGFIX] SQLServer: Send service checks after every run, rather than only at the beginning. See #2515

  • [BUGFIX] vSphere: Enhances topology support, skip unknown metrics. See #2560

  • [BUGFIX] vSphere: The whole check shouldn't fail just because the check failed on a certain instance. See #2548

  • [BUGFIX] Win32: When memory check collection times out, it causes an error in the collector. Instead, it should recover from this. See #2553

  • [BUGFIX] WMI: Allows user to set a profider in request data. See #2565, #2369

5.8.2 / 05-24-2016

Windows only



  • [BUGFIX] Windows: Revert a change introduced in the 5.8.0 release, causing hostnames to change on EC2 hosts when no hostname is defined in datadog.conf, leading to potential backwards compatibility issues. See #2451, #2535

5.8.1 / 05-23-2016

Windows only



  • [BUGFIX] Windows: Fix an issue that was preventing checks from loading properly. See #2528

5.8.0 / 05-23-2016

All platforms


New integrations

  • Kong
  • PowerDNS Recursor
  • Spark

Updated integrations

  • ActiveMQ
  • Cassandra
  • Directory
  • Disk
  • Docker
  • Elasticsearch
  • Ganglia
  • Graphite
  • HAProxy
  • HTTP Check
  • JMX
  • Kubernetes
  • MapReduce
  • Marathon
  • Memcached
  • Mesos Master
  • MongoDB
  • MySQL
  • Process
  • RabbitMQ
  • Riak
  • TeamCity
  • TokuMX
  • Varnish
  • vSphere
  • Windows
  • Windows Service
  • WMI
  • YARN

Service Discovery

Starting with the 5.8 release, the Datadog Agent ships with service discovery. This feature works for Docker containers and can run on platforms such as Kubernetes, Docker Swarm, and Amazon ECS.

Service discovery makes it possible to define configuration templates for a given Docker image in a configuration store, and to dynamically apply them to new containers on the fly. The Agent will then use these configuration templates combined with container metadata to enable, disable and reconfigure checks dynamically in response to container start/stop events.

For more details on Datadog Agent's Service Discovery, check out the wiki page.

See #2008, #2424, #2426, #2514.

Kong integration

Kong is an open-source management layer for APIs.

The checks collects metrics using /status endpoint part of Kong RESTful API. It retrieves usage information about the node, with some basic information about the connections being processed by the underlying NGINX process, and the number of entities stored in the datastore collections. It also submits a service check to reflect the agent's ability to connect to the instance.

See #2241.

A big thank you @shashiranjan84 !

Spark integration

The Spark check runs from the Master Node and retrieves metrics from YARN's ResourceManager. The collected metrics cover statistics about:

  • Spark Jobs
  • Spark Stages
  • Spark Executors
  • Spark Resilient Distributed Datasets (RDD)

See #2407.

PowerDNS Recursor integration

The PowerDNS Recursor check retrieves metrics using the /servers/localhost/statistics endpoint part of PowerDNS API when the experimental web server is active. The collected metrics cover the Recursor statistics: (outgoing) queries, answers, cache and cache entries. A service check reflects the agent's ability to connect to the instance.

For more details about the metrics collected, check out the PowerDNS documentation about Recursor Statistics.

See #2108, #2490.

A big thank you @AntoCard, @janeczku !

MapReduce & YARN updates

Hadoop MapReduce and YARN integrations have been updated to reduce and consolidate the metrics' tags. In particular:

  • cluster_name tag replaces cluster_id
  • app_name tag replaces app_id

MapReduce metrics have received additional improvements and cleanups:

  • Metric type rate replaces gauge when applicable
  • The following metrics have been removed:
    • mapreduce.job.reduce.task.progress

See #2473, #2474, #2487.

Source install: end of Python 2.6 compatibility support

For source installation, the Datadog Agent uses system's Python. Starting with the 5.8.0 release, the Datadog Agent does not enforce compatibility with Python 2.6 anymore and requires Python 2.7. Tests and code sections used to maintain compability have been removed.

Users running a different version of Python on their system are advised to install a packaged version of the Datadog Agent. If you have any questions, please contact our support team support(at)datadoghq(dot)com.

See #2377.


  • [FEATURE] ActiveMQ: Add monitoring for standby hosts. See #2023, #2430 (Thanks @joelvanvelden)

  • [FEATURE] Directory: Add Windows support. See #2410, #2508

  • [FEATURE] Disk: Add an option to exclude disks based on their mountpoints. See #2359, #2476 (Thanks @rhwlo)

  • [FEATURE] Docker: Add optional disk metrics. See #2405

  • [FEATURE] Docker: Add the possibility to tag by container ID. See #2353, #2507

  • [FEATURE] Elasticsearch: Add support for SSL and Elasticsearch Shield. See #2195 (Thanks @pabrahamsson)

  • [FEATURE] Elasticsearch: Collect additional Thread Pool metrics. See #2321

  • [FEATURE] HAProxy: Add option to disable SSL validation. See #2393.

  • [FEATURE] JMX: Add user tags to service checks. See #96

  • [FEATURE] JMX: Allow group name substitutions in attribute/alias parameters. See #94, #97 (Thanks @alz)

  • [FEATURE] Kong: New check. See #2241, Kong integration (Thanks @shashiranjan84)

  • [FEATURE] Kubernetes: Allow to blacklist labels. See #2512

  • [FEATURE] Kubernetes: Tag pods metrics with replica sets. See #2444, #2446, #2512

  • [FEATURE] Memcached: Add "items" and "slabs" statistic metrics as reported by the stats command. See #2491

  • [FEATURE] PowerDNS Recursor: New check. See #2108, #2490, PowerDNS Recursor integration (Thanks @AntoCard, @janeczku)

  • [FEATURE] Process: Add pagefault statistics to process metrics. See #2363, #2477 (Thanks @ovesh)

  • [FEATURE] RabbitMQ: Tag queue based metrics with the queue name as queue_family. See #2259, #2468 (Thanks @rhwlo)

  • [FEATURE] Riak: Collect Riak Search metrics. See #2243 (Thanks @gzysk8)

  • [FEATURE] Service Discovery: Define configuration templates for a given Docker image in a configuration store, and to dynamically apply them to new containers on the fly. See #2008, #2424, #2426, #2514, Service Discovery

  • [FEATURE] Spark: New check. See #2407, #2479, Spark integration

  • [FEATURE] TeamCity: Support basic HTTP authentication. See #2266 (Thanks @mderomph-coolblue)

  • [FEATURE] vSphere: Add the ability to set a custom field from the vSphere PowerCLI to mark the VMs to monitor. See #2459

  • [FEATURE] Windows Event Log: Add an optional event_format parameter to instruct the check to generate Datadog's event bodies with the specified list of event properties. See #2441

  • [FEATURE] Windows: Add opened handles statistics to system metrics. See #2403

  • [FEATURE] Windows: Add swap -pagefile- statistics to system metrics. See #2422

  • [IMPROVEMENT] Core: Allow AgentCheckTest facilities to be used on SDK check modules. See #2341

  • [IMPROVEMENT] Core: Gracefully handle responses with HTTP status code 413. See #2418

  • [IMPROVEMENT] Core: Update the configuration logic to look for checks in a 3rd-party directory. See #2273, #2467

  • [IMPROVEMENT] Directory: Speed up the file count in a directory. See #2245 (Thanks @tebriel)

  • [IMPROVEMENT] Gohai: Collect process resources and deprecate the corresponding Python check. See #24, #2218

  • [IMPROVEMENT] Gohai: Improve logging. See #30

  • [IMPROVEMENT] Kubernetes: Enable kubelet check by default. See #2512

  • [IMPROVEMENT] Kubernetes: Improve support for Kubernetes >= 1.2. See #2388, #2494, #2512

  • [IMPROVEMENT] Kubernetes: Timeout on slow requests. See #2345

  • [IMPROVEMENT] MapReduce: Consolidate and reduce the number of tags produced. See #2474, MapReduce & YARN updates

  • [IMPROVEMENT] MapReduce: Switch metric types from gauge to rate when applicable. See #2487, MapReduce & YARN updates

  • [IMPROVEMENT] Marathon: Remove versions metric to decrease the check collection runtime. See #1861, #2443 (Thanks @Zarkantho)

  • [IMPROVEMENT] MongoDB: Add additional tag on status change to make searching more precise. See #2502 (Thanks @gphat)

  • [IMPROVEMENT] MongoDB: Improve metric tagging. See #2457

  • [IMPROVEMENT] MySQL: Improve accuracy for query run time averages. See #2406, #2463

  • [IMPROVEMENT] YARN: Consolidate and reduce the number of tags produced. See #2473, MapReduce & YARN updates

  • [BUGFIX] Core: Fix DogStatsd's service check parsing when a tag key ends with m. See #2373, #2472

  • [BUGFIX] Core: Restart the forwarder when a "pathological" activity is detected. See #2322

  • [BUGFIX] Core: Retrieve EC2 instance ID on Windows. See #2451

  • [BUGFIX] Docker: Fix container ID extraction for RancherOS. See #2454, #2466 (Thanks @ricky26)

  • [BUGFIX] Docker: Fix improper logging causing the check to raise an exception when excluding containers. See #2464 (Thanks @yenif)

  • [BUGFIX] Docker: Ignore mountpoints not accessible by Datadog Agent dd-agent user. See #2427 (Thanks @c960657)

  • [BUGFIX] Elasticsearch: Restrict the list of metrics to collect by version. See #2186, #2232 (Thanks @bdharrington7)

  • [BUGFIX] HTTP Check: Support no_proxy environment variable. See #2448

  • [BUGFIX] JMX: Report properly beans with ':' in the name. See #90, #91, #95 (Thanks @bluestix)

  • [BUGFIX] JMX: Sanitize metric names and tags, i.e. remove illegal characters. See #89

  • [BUGFIX] JMX: Support attribute types. See #92 (Thanks @nwillems)

  • [BUGFIX] JMX: Use localhost interface when StatsD is bound to all interfaces. See #2516

  • [BUGFIX] Kubernetes: Handle case where no pods are running. See #2263, #2512

  • [BUGFIX] Mesos Master: Fix key error when a metric is missing. See #2500 (Thanks @gphat)

  • [BUGFIX] MongoDB: Attach the check's events with the integration in Datadog. See #2456

  • [BUGFIX] Packaging: Exclude an operating-system-specific .dll from the MSI installer that was causing installation and runtime failures on Windows Server 2008. See #2486

  • [BUGFIX] TokuMX: Fix AttributeError exceptions when collecting metrics from a replica. See #2390

  • [BUGFIX] TokuMX: Fix an issue among metrics collection causing rate metric values to be potentially wrong and ZeroDivisionError exceptions. See #2510

  • [BUGFIX] TokuMX: Support metric values with Int64 type. See #2390

  • [BUGFIX] Windows Service: Make Windows service names case-insensitive. See #2492

  • [OTHER] Cassandra: Remove deprecated options from the datadog.conf configuration file. See #2447

  • [OTHER] Core: Remove support of Python 2.6. See #2377, Source install: end of Python 2.6 compatibility support

  • [OTHER] Core: Upgrade boto dependency from 2.36.0 to 2.39.0. See #2402, #3399

  • [OTHER] Disk: Remove deprecated options from the datadog.conf configuration file. See #2447

  • [OTHER] Ganglia: Remove deprecated options from the datadog.conf configuration file. See #2447

  • [OTHER] Graphite: Remove deprecated options from the datadog.conf configuration file. See #2447

  • [OTHER] HTTP Check: Remove the deprecated threshold and window parameters from the configuration file. See #2400

  • [OTHER] Memcached: Document usage for listening to a Unix socket. See #2298, #2431 (Thanks @kzw)

  • [OTHER] TokuMX: Document usage for connecting to a MongoDB instance with a "full" URI. See #2390

  • [OTHER] Varnish: Document varnish group requirement. See #2384

  • [OTHER] WMI: Document wildcard usage in filters. See #2368, #2469 (Thanks @mausch)

5.7.4 / 04-21-2016

All platforms



  • [FEATURE] Core: Add Python architecture to the info command. See #2413

  • [FEATURE] MongoDB: Collect additional WiredTiger storage engine metrics. See #1825, #2423 (Thanks @benmccann)

  • [FEATURE] MongoDB: Collect replication set information metrics. See #2237, #2429 (Thanks @rhwlo)

  • [IMPROVEMENT] Cassandra: Exclude new system_schema keyspace in version 3.x. See #2339

  • [IMPROVEMENT] Kafka: Update the default configuration file to support collection of more metrics. See #2371, #2425

  • [IMPROVEMENT] MongoDB: Collect journaling, locks, WiredTiger storage engine metrics by default and deprecate the corresponding options. See #2423

  • [IMPROVEMENT] MongoDB: Improve messaging and tags on replication set member state events. See #2409, #2432 (Thanks @antifuchs)

  • [IMPROVEMENT] Windows WMI-based checks (wmi_check, System check, IIS, Windows Service, Windows Event Log): gracefully time out WMI queries. See #2185, #2228, #2278, #2366 and #2401.

  • [BUGFIX] Agent Metrics: Flush service metadata to avoid memory leaks. See #2414

5.7.3 / 03-31-2016

All platforms



  • [IMPROVEMENT] Linux install script: Ignore apt-get update failures and use https for apt repo. See #2378

  • [IMPROVEMENT] WMI check: Make configuration of the metric types case-insensitive. See #2392

  • [BUGFIX] Consul: Enforce that get_peers_in_cluster returns a list. See #2381

  • [BUGFIX] Core: Fix Japanese tzname encoding issue on Windows. See #2351

  • [BUGFIX] Core: On RHEL, make the stop init command kill all the Agent processes properly. See #2349

  • [BUGFIX] Core: Fix Watchdog timeout duration on the forwarder. See #2320

  • [BUGFIX] IIS: Fix CRITICAL service check on the _Total site (e.g. when no sites are specified). See #2387

  • [BUGFIX] IIS: Deal with BytesTransfered vs BytesTransferred 2008sp2 typo. See #2379

  • [BUGFIX] JMXFetch: Take into account bind_host. See #2372 and jmxfetch-85

  • [BUGFIX] JMXFetch: Handle IOException gracefully at the instance level. See jmxfetch-83

  • [BUGFIX] Packaging: Fix the version of requests shipped with the packaged Linux Agent. See omnibus-software-45

  • [BUGFIX] MySQL: Avoid check failure when InnoDB is not available or disabled. See #2385

  • [BUGFIX] SNMP: Fix errors in multiple-instance configurations caused by thread-safety issues with pysnmp cmd_generator. See #2357

  • [BUGFIX] SQLServer: Fix connection to DB when no username or password are specified. See #2311

  • [BUGFIX] vSphere: Fix SSL config options feature by upgrading the packaged version of pyvmomi. See omnibus-software-44

  • [BUGFIX] Windows Event Log: Fix check when tag_event_id:true #2397

5.7.2 / 03-17-2016

Windows only



  • [BUGFIX] WMI: Disable query timeout, cache and re-use connections to avoid memory leaks. See #2366

5.7.1 / 03-09-2016

All platforms

NB: For Windows, please also refer to the 5.7.0 section. 5.7.0 was not released on Windows but the 5.7.1 Windows installer includes all the changes listed in the 5.7.0 section.



  • [BUGFIX] Core: Avoid python segfault when the ctypes module is imported on SELinux-enabled environments. See omnibus-software-43
  • [BUGFIX] MySQL: Fix check failure when no tag is provided. See #2329
  • [BUGFIX] Packaging: Fix RPM package for Amazon Linux EMR. See omnibus-ruby-18

5.7.0 / 03-07-2016

Linux, Mac OS and Source Install only


New integrations

  • Ceph
  • DNS
  • HDFS
  • MapReduce
  • StatsD
  • TCP RTT (go-metro)
  • YARN

Updated integrations

  • Apache
  • AWS
  • Cassandra
  • Consul
  • Directory
  • Docker
  • Elasticsearch
  • Go expvar
  • Gunicorn
  • HAProxy
  • HTTP
  • IIS
  • Kafka
  • Mesos
  • MongoDB
  • MySQL
  • PgBouncer
  • Postgres
  • Process
  • Redis
  • SNMP
  • SSH
  • TeamCity
  • Tomcat
  • vSphere
  • Windows Service
  • Windows Event Log
  • WMI
  • Zookeeper

Hadoop integrations (HDFS, MapReduce and YARN checks)

The Agent now includes 4 new checks to monitor Hadoop clusters:

  • 2 HDFS checks (hdfs_namenode and hdfs_datanode) that collect metrics respectively from namenodes and datanodes using the JMX-HTTP API
  • a MapReduce check that collects metrics on the running Mapreduce jobs from the Application Master's REST API
  • a YARN check that collects metrics from YARN's ResourceManager REST API

The existing hdfs check is deprecated and will be removed in a future version of the Agent. Its metric scope is entirely covered by the new hdfs_namenode check.

TCP RTT measurement with go-metro

This new feature is in beta

The Datadog Agent on 64-bit Linux is now bundled with a new component (go-metro) that passively calculates TCP RTT metrics between the agent's host and external hosts, and reports them as, and through StatsD.

go-metro follows TCP streams active within a certain period of time and estimates the RTT between any outgoing packet with data, and its corresponding TCP acknowledgement.

go-metro runs in its own process. It's disabled by default and can be enabled like a regular check by configuring an /etc/dd-agent/conf.d/go-metro.yaml file and restarting the agent.

For more details on go-metro, check out the project's GitHub page.

Ceph check

The Ceph check retrieves metrics from Ceph's Administration Tool command (ceph).

The check collects metrics from mon_status, status, df detail, osd pool stats and osd perf, and sends a service check reflecting the overall health of the cluster.

See #2264


Multiple community-contributed additions to the MySQL check have been consolidated and merged, including:

  • metrics from the performance_schema table on MySQL >= 5.6 (thanks to @ovaistariq)
  • extra metrics on the InnoDB and MyISAM engines, from the Binlog, and from the SHOW STATUS query (thanks to @ovaistariq)
  • several schema-specific metrics, including schema size, schema average query runtime and 95th percentile query execution time (thanks again to @ovaistariq)
  • metrics on the Handler (thanks to @polynomial)
  • Galera-specific performance stats (thanks to @zdannar)
  • Query Cache metrics (thanks to @leucos)
  • a mysql.replication.slave_running service check reflecting the state of the slaves (thanks to @c960657)

Most of these additional metrics are not collected by default but can be enabled in the check's YAML file. See the YAML conf example file for details.

Various bug fixes and improvements have also been implemented:

  • the Agent's connections to MySQL are handled properly to prevent stale connections
  • the replication status is implemented on both the master and the slaves. On the master this status is determined by the Binlog status and the number of slaves.
  • the system metrics of MySQL are retrieved w/o errors on non-linux platforms by using the psutil library
  • the parsing of the MySQL server version is improved

Huge thanks to all our contributors for all these improvements!

See #2116 and #2242

Potential backward incompatibilities


The dockerized Agent now uses the docker hostname (provided by the Name param from docker info) as its own hostname when available. This means that for hosts running the dockerized Agent the reported hostname may change to this docker-provided hostname.

For reference, the rules followed by the Agent for its hostname resolution are described on this wiki page.


The collect_tcmalloc_metrics parameter in the YAML conf is replaced with the tcmalloc option under additional_metrics. Please refer to the example YAML conf file for more info on the usage of the additional_metrics option.


Instead of sending all metrics as gauges, the vSphere integration now checks the types of the metrics as reported by the VMWare module, and sends metrics as rates when applicable.

If you haven't enabled the all_metrics option on the check, the only affected metrics are cpu.usage, cpu.usagemhz, network.received and network.transmitted. If the option is enabled, the additional affected metrics are listed here. The change will affect the values of these metrics.

WMI check

The wmi_check now only supports % as the wildcard character in the filters. The support of * as the wildcard character, which was undocumented, has been dropped.


  • [FEATURE] Ceph: New check collecting metrics from Ceph clusters. #2264

  • [FEATURE] Consul: Add SSL support. See #2034 (Thanks @diogokiss)

  • [FEATURE] DNS: New check that sends a service check reflecting the status of a hostname's resolution on a nameserver. See #2249 and #2289

  • [FEATURE] Elasticsearch: Report additional metrics related to fs, indices.segments and indices.translog. See #2143 (Thanks @bdharrington7)

  • [FEATURE] HDFS: 2 new checks (see description above). See #2235, #2260, #2274 and #2287

  • [FEATURE] Go-metro: New component that measures TCP RTT (in beta, see description above). See #2208

  • [FEATURE] Linux: Add memory metrics (slab, page tables and cached swap). See #2100 (Thanks @gphat)

  • [FEATURE] Linux: New linux_proc_extras check collecting system-wide metrics on interrupts, context switches and processes. See #2202 (Thanks @gphat)

  • [FEATURE] MapReduce: New check (see description above). See #2236

  • [FEATURE] MongoDB: Collect optional additional metrics, grouped by topic. These can be enabled with the new additional_metrics option in the check's YAML conf. Also, the underlying pymongo library has been upgraded from 2.8 to 3.2. See #2161, #2166, #2140 and #2160 (Thanks @scottbessler and @benmccann)

  • [FEATURE] MySQL: Add tag parameter for custom MySQL queries. See #2229

  • [FEATURE] MySQL: Enhance the catalog of metrics reported, and add a service check on the replication state. See #2116, #2242 and #2288 (Thanks @ovaistariq, @zdannar, @polynomial, @leucos, @Zenexer, @c960657, @nfo, @patricknelson and @scottgeary)

  • [FEATURE] Postgres: Measure user functions. See #2164

  • [FEATURE] Process: Allow configuring the path to procfs (useful when the agent is run in a container), with a newer version of psutil. See #2163 and #2134 (Thanks @sethp-jive)

  • [FEATURE] Redis: Optionally report metrics from INFO COMMANDSTATS as calls, usec and usec_per_call (prefixed with redis.command.). See #2109

  • [FEATURE] SNMP: Add support for forced SNMP data types to help w/ buggy devices. See #2165 (Thanks @chrissnell)

  • [FEATURE] SSH: Add Windows support. See #2072

  • [FEATURE] StatsD: New check collecting metrics and service checks using StatsD's admin interface. See #1978 and #2162 (Thanks @gphat)

  • [FEATURE] vSphere: Add SSL config options for certs. See #2180

  • [FEATURE] YARN: New check (see description above). See #2147 and #2207

  • [FEATURE] Zookeeper: Gather stats from mntr command and report zookeeper.instances.<mode> metrics as 0/1 gauge. See #2156 (Thanks @jpittis)

  • [IMPROVEMENT] Apache: Allow disabling ssl validation. See #2169

  • [IMPROVEMENT] AWS: Incorporate security-groups into tags collected from EC2. See #1951

  • [IMPROVEMENT] Cassandra: Add YAML conf for Cassandra version > 2.2. See #2142 and #2271

  • [IMPROVEMENT] Directory: Show check on Windows. See #2184 (Thanks @xkrt)

  • [IMPROVEMENT] Docker: Pass tags to events as well. See #2182

  • [IMPROVEMENT] Docker: Use the docker hostname as the agent's hostname when available. See #2145

  • [IMPROVEMENT] Elasticsearch: Apply custom tags to service checks too. See #2148

  • [IMPROVEMENT] Go expvar: Add configuration option for custom metric namespace. See #2022 (Thanks @theckman)

  • [IMPROVEMENT] Go expvar: Add counter support. See #2133 (Thanks @gphat)

  • [IMPROVEMENT] Gohai: Count number of logical processors. See gohai-22

  • [IMPROVEMENT] HAProxy: Add option to count statuses by service. See #2304 and #2314

  • [IMPROVEMENT] HTTP: Add a days_critical option to the SSL certificate expiration feature. See #2087

  • [IMPROVEMENT] HTTP: Support unicode in content-matching. See #2092

  • [IMPROVEMENT] Kafka: Compute instant rates and capture more metrics in example configuration. See #2079 (Thanks @dougbarth)

  • [IMPROVEMENT] Linux install script: Add custom provided hostname to datadog.conf. See #2225 (Thanks @lowl4tency)

  • [IMPROVEMENT] Mesos: Improve checks' performance by preventing requests from using chardet. See #2192 (Thanks @GregBowyer)

  • [IMPROVEMENT] MongoDB: Tag mongo instances by replset state. See #2244 (Thanks @rhwlo)

  • [IMPROVEMENT] SNMP: Improve performance by running instances of the check in parallel. See #2152

  • [IMPROVEMENT] SNMP: Make MIB constraint enforcement optional and improve resilience. See #2268

  • [IMPROVEMENT] TeamCity: Allow disabling ssl validation. See #2091 (Thanks @jslatts)

  • [IMPROVEMENT] Unix: Revamp source install script. See #2198 and #2199

  • [IMPROVEMENT] vSphere: Add network.received and network.transmitted to the basic metrics collected. See #1824

  • [IMPROVEMENT] vSphere: Check metric type to determine how to report (rate or gauge). See #2115

  • [IMPROVEMENT] Windows: Add uptime metric. See #2135, #2292 and #2299

  • [IMPROVEMENT] Windows WMI-based checks (wmi_check, System check, IIS, Windows Service, Windows Event Log): gracefully time out WMI queries. See #2185, #2228 and #2278

  • [IMPROVEMENT] Windows IIS, Service and Event Log checks: use the new WMI wrapper with increased performance. See #2136

  • [IMPROVEMENT] Windows packaging: Tighten permissions on datadog.conf. See #2210

  • [BUGFIX] AWS: Use proxy settings for EC2 tag collection. See #2201

  • [BUGFIX] AWS: During EC2 tags collection, log a warning when the instance is not associated with an IAM role. See #2285

  • [BUGFIX] Core: Do not log API keys. See #2146

  • [BUGFIX] Core: Fix cases of low/no disk space causing the Agent to crash when calling subprocesses. See #2223

  • [BUGFIX] Core: Make Dogstatsd recover gracefully from serialization errors. See #2176

  • [BUGFIX] Core: Set agent pid file and path from constants. See #2128 (Thanks @urosgruber)

  • [BUGFIX] Core: Log check starts with info level on first run only. See #2772

  • [BUGFIX] Development: Fix test of platform in etcd CI setup script. See #2205 (Thanks @ojongerius)

  • [BUGFIX] Docker: Avoid event collection failure if an event has no ID param. See #2308

  • [BUGFIX] Docker: Catch exception when getting k8s labels fails. See #2200

  • [BUGFIX] Docker: Don't warn if process finishes before measuring. See #2114 (Thanks @oeuftete)

  • [BUGFIX] Docker: Remove misleading warning on excluded containers. See #2179 (Thanks @EdRow)

  • [BUGFIX] Documentation: Update link to dogstatsd guide in datadog.conf. See #2181

  • [BUGFIX] Elasticsearch: Optionally collect pending task stats. See #2250

  • [BUGFIX] Flare: Use ssl and proxy settings from datadog.conf. See #2234 (Thanks @tebriel)

  • [BUGFIX] Flare: Mention path to tar file in Windows UI. See #2084

  • [BUGFIX] FreeBSD: Use correct log file for syslog. See #2171

  • [BUGFIX] Go expvar: Add timeout for requests to get go expvar metrics. See #2183 (Thanks @gphat)

  • [BUGFIX] Gohai: Log unexpected OSError exceptions instead of re-raising them. See #2309

  • [BUGFIX] Gunicorn: Mention in YAML conf that the setproctitle module is required. See #2215

  • [BUGFIX] HTTP: Add an option to disable warnings when ssl validation is disabled. See #2193

  • [BUGFIX] HTTP: Improve log message when http code is incorrect. See #2203

  • [BUGFIX] HTTP: Rename ssl_expire to check_certificate_expiration in YAML comment. See #2086 (Thanks @MiguelMoll)

  • [BUGFIX] HTTP: Use proxy settings from datadog.conf. See #2112

  • [BUGFIX] Kubernetes: Remove unused function. See #2157

  • [BUGFIX] OpenStack: Improve docs in YAML conf. See #2094

  • [BUGFIX] OpenStack: Remove recommendation for omitting trailing slashes in YAML conf. See #2081

  • [BUGFIX] Mac OS X: Fix gohai call by passing correct PATH to supervisor. See #2206

  • [BUGFIX] Mesos slave: Allow configuring mesos master port. See #2189

  • [BUGFIX] MySQL: Fix buggy tagging in service_checks on instances configured w/ unix socket. See #2216

  • [BUGFIX] PgBouncer: Avoid raising error when there are no results for a query. See #2280 (Thanks @hjkatz)

  • [BUGFIX] SNMP: Fix bug when the requested oid is prefixed by another requested oid. See #2246 (Thanks @xkrt)

  • [BUGFIX] Tomcat: Fix bad attribute in YAML conf file. See #2153

  • [BUGFIX] Unix: Fix URL of get-pip script in source install script. See #2220 (Thanks @mooney6023)

  • [BUGFIX] Windows: Fix cases of collector getting wrongfully restarted by watchdog after one correct watchdog restart. See #2175

  • [BUGFIX] WMI check: Remove unnecessary warnings on Name property. See #2291

  • [BUGFIX] WMI check: Always add the tag_by parameter to the collected properties. See #2296

5.6.3 / 12-10-2015



  • [FEATURE] Consul: More accurate nodes_* and services_* gauges (NB: consul.check service checks are now tagged by consul_service_id rather than service-id) See #2130 (Thanks @mtougeron)

  • [FEATURE] Docker: Improve container name, image name and image tag extraction. See #2071

  • [BUGFIX] Core: Catch and log exceptions from the resources checks. See #2029

  • [BUGFIX] Core: Fix host tags sending when create_dd_check_tags is enabled. See #2088

  • [BUGFIX] Docker: Add one more cgroup location. See #2139 (Thanks @bakins)

  • [BUGFIX] Flare: Remove proxy credentials from collected datadog.conf. See #1942

  • [BUGFIX] Marathon: Fix disk typo in metric name. See #2126 (Thanks @pidah)

  • [BUGFIX] OS X: Fix memory metrics. See #2097

  • [BUGFIX] Postgres: Fix metrics not reporting with multiple relations. See #2111

  • [BUGFIX] Windows: Bundle default CA certs of requests. See #2098

5.6.2 / 11-16-2015

Linux, Mac OS and Source Install only



  • [FEATURE] Docker/Kubernetes: Collect Kubernetes labels as tags. See #2075, #2082

  • [FEATURE] HTTPCheck: Option to support -RSA, RC4, MD5- weak SSL/TLS ciphers. See #1975, #2048

  • [BUGFIX] Core: Improve detection of agent process from PID to avoid false positives. See #2005

5.6.1 / 11-09-2015



  • [BUGFIX] Consul: Add the main tags to service checks. See #2015 (Thanks @mtougeron)

  • [BUGFIX] Docker: Remove spurious proc root container warnings. See #2055 (Thanks @oeuftete)

  • [BUGFIX] Flare: Restore missing JMXFetch information. See #2062

  • [BUGFIX] OpenStack: Fix false-critical on the network service check. See #2063

  • [BUGFIX] Windows: Restore missing JMXFetch service logs. See #1852, #2065

  • [OTHER] Upgrade pymongo dependency from 2.6.3 to 2.8 on Windows Datadog Agent 32-bit MSI Installer.

  • [OTHER] Allow supervisor.conf to select Supervisor user. See #2064

5.6.0 / 11-05-2015

Linux, Mac OS and Source Install only


New integration(s)

  • Kubernetes
  • OpenStack

Updated integrations

  • ActiveMQ
  • Cassandra
  • Couchbase
  • Docker
  • Dogstream
  • HAProxy
  • HTTPCheck
  • JMXFetch
  • Memcached
  • MongoDB
  • Network
  • Nginx
  • Process
  • Riak
  • SNMP
  • SQL Server
  • Unix
  • Windows
  • Windows Event Viewer
  • WMI

Kubernetes check

The Kubernetes check retrieves metrics from cAdvisor running under Kubelet.

See #2031

OpenStack check

The OpenStack check is intended to run besides individual hypervisors. It can be scoped to any set of projects living on that host via instance-level config.

At the hypervisor-level it collects:

At the project-level it collects:

Additionally it sends service checks to register the UP/DOWN state of networks discovered by the agent, the locally running hypervisor, and the outward-facing (public || internal) API services of Nova, Neutron and Keystone


While the check will run without issues as an admin user, it is recommended to configure a read-only datadog user and configure the check with the corresponding user/password. Instructions on setting up the datadog user + role , as well as the changes required to the policy.json file can be found here

A note on compatibility

Authentication is performed via the password method, and requires Identity API v3 Nova API v2 and v2.1 are supported, with minor additional configuration necessary for v2.

A big thank you @mtougeron !

See #1864

New WMI module wrapper

Datadog Agent 5.6.0 ships a new built-in lightweight Python WMI module wrapper, built on top of pywin32 and win32com extensions.


  • Based on top of the pywin32 and win32com third party extensions only
  • Compatible with Raw* and Formatted Performance Data classes
    • Dynamically resolve properties' counter types
    • Hold the previous/current Raw samples to compute/format new values*
  • Fast and lightweight
    • Avoid queries overhead
    • Cache connections and qualifiers
    • Use wbemFlagForwardOnly flag to improve enumeration/memory performance

* Raw data formatting relies on the avaibility of the corresponding calculator. Please refer to checks.lib.wmi.counter_type for more information

The new WMI module wrapper is used among the following checks to improve speed performances:

  • System
  • WMI

Other checks relying on WMI collection will follow in future versions of Datadog Agent.

See #2011
Original discussion thread: #1952
Credits to @TheCloudlessSky (

[Warning] JMXFetch false-positive bean match & potential backward incompatibilities issues

JMXFetch was illegitimately matching some MBeans attributes when the associated MBean had one of its parameter defined in an instance configuration.

The issue is addressed. As a result, please note that metrics related to false positive bean matches are not reported anymore.

Potential affected checks: ActiveMQ, Cassandra, JMX, Solr, Tomcat.

For more information, please get in touch with support(at)datadoghq(dot)com

See #81


  • [FEATURE] Cassandra: Support Cassandra > 2.2 metric name structure (CASSANDRA-4009). See #79, #2035

  • [FEATURE] Core: Add service check count to the output of Dogstatsd 'info' section. See #1799

  • [FEATURE] Docker: Add container names as tags for events. See #2026

  • [FEATURE] HAProxy: Collect the number of available/unavailable backends. See #1915 (Thanks @a20012251)

  • [FEATURE] JMXFetch: Option to add custom JARs to the classpath. See #1996

  • [FEATURE] JMXFetch: Support float and java.lang.Float attribute types as simple JMX attributes. See #76

  • [FEATURE] JMXFetch: Support Cassandra > 2.2 metric name structure (CASSANDRA-4009). See #79

  • [FEATURE] JMXFetch: Support custom JMX Service URL to connect to, on a per-instance basis. See #80

  • [FEATURE] Kubernetes: New check. See #1919, #2031, #2038, #2039

  • [FEATURE] Memcached: Collect listen_disabled_num timeout counter. See #1995 (Thanks @alaz)

  • [FEATURE] MongoDB: Collect TCMalloc memory allocator metrics. See #1979 (Thanks @@benmccann)

  • [FEATURE] MongoDB: Report dbStats metrics for all databases. See #1855, #1961 (Thanks @asiebert)

  • [FEATURE] Network: Add UDP metrics from /proc/net/snmp in addition to the existing TCP metrics. See #1974, #1986 (Thanks @gphat)

  • [FEATURE] OpenStack: New check. See #1864, #2040

  • [FEATURE] Riak: Add custom tags to service checks' tags. See #1482, #1527, #1987

  • [FEATURE] SNMP: Option to set the OID batch size. See #1990

  • [FEATURE] Unix: Collect /proc/meminfo MemAvailable metric when available. See #1826, #1993 (Thanks @jraede)

  • [FEATURE] Windows Event Viewer: Option to tag events by event_id. See #2009

  • [IMPROVEMENT] Core: Deprecate 'use_dd' flag. See #1856, #1860 (Thanks @ssbarnea)

  • [IMPROVEMENT] Core: Fix hanging subprocess.Popen calls caused by buffer limits. See #1892

  • [IMPROVEMENT] Core: Remove the uses of list comprehensions as looping constructs. See #1939 (Thanks @jamesandariese)

  • [IMPROVEMENT] Core: Run Supervisor as dd-agent user. See #1348, #1620, #1895

  • [IMPROVEMENT] Core: Use user-defined NTP settings in 'info' command's status page. See #1985

  • [IMPROVEMENT] Dogstream: Add DEBUG logging to event collection. See #1910

  • [IMPROVEMENT] JMXFetch: Assign generic alias if not defined. See #78

  • [IMPROVEMENT] Network: Use ss instead of netstat on Linux systems. See #1156, #1859 (Thanks @tliakos)

  • [IMPROVEMENT] Nginx: Add logging on check exceptions. See #1813, #1914 (Thanks @clokep)

  • [IMPROVEMENT] Process: Improve sampling of system.processes.cpu.pct metric. See #1660, #1928

  • [IMPROVEMENT] Unix: Filter SunOS memory_cap kstats by module. See #1959 (Thanks @pfmooney)

  • [IMPROVEMENT] Windows: New WMI module wrapper to improve speed performances. See #1952, #2011 (Thanks @TheCloudlessSky)

  • [IMPROVEMENT] Windows: Switch to the built-in WMI core to improve system metric collection performances. See #1952, #2011 (Thanks @TheCloudlessSky)

  • [IMPROVEMENT] WMI: Switch to the built-in WMI core to improve the check performances. See #1952, #2011 (Thanks @TheCloudlessSky)

  • [BUGFIX] ActiveMQ: Limit metric collection to the queues specified in the configuration file. See #1948 (Thanks @joelvanvelden)

  • [BUGFIX] Core: Normalize exit code of fallback status command. See #1976, #1988 (Thanks @wyaeld)

  • [BUGFIX] Couchbase: Only send selected bucket-level metrics. See #1936

  • [BUGFIX] Docker: Do not run on initialization failures. See #1984

  • [BUGFIX] Docker: Improve container name extraction to avoid duplicates. See #1965

  • [BUGFIX] Flare: Obfuscate passwords encoded in URIs. See #2010

  • [BUGFIX] HTTPCheck: Fix SSL Certificate check when specifying a port in the URL. See #1923, #1944 (Thanks @dmulter)

  • [BUGFIX] JMXFetch: Fix bean name matching logic: ORAND. See #81

  • [BUGFIX] Process: Avoid refreshing AccessDenied PIDs cache at every run. See #1928

  • [BUGFIX] SQL Server: Close database connections so SQL Server Agent can stop. See #1997

  • [BUGFIX] Windows: Limit Datadog Agent Manager to a single instance. See #1924, #1933

5.5.2 / 10-26-2015


[WARNING] Datadog Agent not reporting metrics after Daylight Saving Time (DST) ends

This release fixes a bug on servers that are configured in local time instead of UTC Time. If your server's clock is configured to use Daylight Saving Time (DST), the Datadog Agent might stop sending metrics for up to one hour when the Daylight Saving Time ends or until it is restarted after the Daylight Saving Time ends.

We highly recommend to upgrade to this version if your server is configured in local time.


  • [BUGFIX] Consul: Send the health state service checks of all nodes. See #1900 (Thanks @jgmchan)
  • [BUGFIX] Core: Use utcnow instead of now to avoid the forwarder to run into a locked state. See #2000
  • [BUGFIX] Fix pycurl dependency issue with Windows Datadog Agent 64-bit MSI Installer.

5.5.1 / 09-23-2015



  • [BUGFIX] Core: Fix dd-agent command-line interface on Linux. See #49, #51

  • [BUGFIX] Docker: Fix Amazon EC2 Container Service (ECS) tags collection. See #1932

  • [BUGFIX] Docker: Improve parsing of the cpuacct field and of the container ID. See #1940 (Thanks @joshk0)

  • [BUGFIX] HTTP Check: Fix SSL certificate check when specifying a non-default port in the URL. See #1923 (Thanks @dmulter)

  • [BUGFIX] Nginx: Fix 'application/json' content_type support. See #1943

  • [OTHER] Windows: Ship latest version of Gohai with Windows MSI Installer.

5.5.0 / 09-17-2015


New integration(s)

  • Consul

Updated integrations

  • Agent Metrics
  • Amazon EC2
  • Btrfs
  • Couchbase
  • CouchDB
  • Disk
  • Docker
  • Elasticsearch
  • etcd
  • Google Compute Engine
  • HTTP Check
  • JMXFetch
  • Mesos
  • MongoDB
  • MySQL
  • Network
  • Nginx
  • PgBouncer
  • PostgreSQL
  • Process
  • RabbitMQ
  • Redis
  • Supervisor
  • System
  • Unix
  • Windows Event Viewer
  • WMI

Consul check

New Consul check.

Supported metrics:

  • Number of Consul Agents in the Cluster consul.peers: tagged by consul_datacenter and mode (leader | follower)
  • Consul Catalog Nodes Up by Service consul.catalog.nodes_up: tagged by consul_datacenter and consul_service_id
  • Consul Catalog Services Up by Node consul.catalog.services_up: tagged by consul_datacenter and consul_node_id

Supported events:

  • consul.new_leader events when a leader change is detected.

See #1628

New Docker check

Datadog agent 5.5.0 introduces a new Docker check: 'docker_daemon'.

In terms of features, it adds:

  • Support for TLS connections to the daemon
  • New metrics:
  • Network metrics
  • Memory limits
  • Container size (rootfs)
  • Image size
  • Support for labels (convert them into tags). Off by default, uses a list of labels that should be converted.
  • Support for ECS tags: task name and task version

Backward incompatible changes:

As a consequence, the previous check 'Docker' is now deprecated and will not receive further support.

See #1908

Windows 64bit - Datadog Agent

The Datadog Agent is now available in a 64bit version on Windows. For more information, please visit our Integrations/Agent page.

Flare on Windows

Datadog Agent flare feature makes easy to ship a tarball with logs and configurations to ease agent troubleshooting. Previously exclusive to Linux, it's now available on Windows. For more information, please visit our wiki page.

[Warning] JMX host tag issues & potential backward incompatibilities issues with service check monitors

JMX related checks -c.f. list below- were illegitimately submitting service checks tagged with the host value defined in the YAML configuration file. As it was overriding the agent hostname, with a value often equals to localhost, it was difficult to define and scope monitors based on these service checks.

The issue is addressed. JMX service checks have a new jmx_server tag which contains the YAML configuration host value so it does not replace the actual agent hostname in the host tag anymore.

Warning: these changes affect your JMX-service-checks related existing monitors scoped with the host tag. For more information, please get in touch with

JMX related checks: ActiveMQ, Cassandra, JMX, Solr, Tomcat.

See #66

Deprecation notice

datadog.conf disk options

Disk options in datadog.conf file are being deprecated to promote the new Disk check introduced in the 5.4.0 release. It will be removed in a future version of the Datadog Agent. Please consider conf.d/disk.yaml instead to configure it.

See #1758

Generic Mesosphere check

The previous generic Mesosphere check is deprecated, in favor of the Mesosphere master and slave specific checks introduced in the 5.4.0 release. It will be removed in a future version of the Datadog Agent.

See #1535

Previous Docker check

The previous Docker check is deprecated, in favor of the new one introduced in the 5.5.0 release. It will be removed in a future version of the Datadog Agent.

See #1908


  • [FEATURE] Consul: New check reporting cluster, service and node wide metrics and events for leader election. See #1628

  • [FEATURE] CouchDB: Allow blacklisting of specific databases. See #1760

  • [FEATURE] Docker: New Docker check. See #1908

  • [FEATURE] Elasticsearch: Collect common JVM metrics. See #1865

  • [FEATURE] Elasticsearch: Collect primary shard statistic metrics. See #1875

  • [FEATURE] etcd: SSL support. See #1745 (Thanks @KnownSubset)

  • [FEATURE] Flare: Add JMXFetch-specific information. See #1726

  • [FEATURE] Flare: Log permissions on collected files. See #1767

  • [FEATURE] Flare: Windows support. See #1773

  • [FEATURE] HTTP Check: Add SSL certificate configuration and validation options. See #1720

  • [FEATURE] JMXFetch: Memory saving by limiting MBeans queries to certain scopes. See #63

  • [FEATURE] JMXFetch: Wildcard support on domains and bean names. See #57

  • [FEATURE] MongoDB: Collect active client connections metrics. Enhance connections, dbStats, mem and rpl metric coverage. See #1798

  • [FEATURE] MongoDB: Make timeout configurable and increase the default. See #1823 (Thanks @benmccann)

  • [FEATURE] MySQL: Custom query metrics. See #1793 (Thanks @obi11235)

  • [FEATURE] Nginx: Option to disable SSL validation. See #1626 #1782

  • [FEATURE] PostgreSQL: SSL support. See #1696 (Thanks @bdotdub)

  • [FEATURE] PostgreSQL: Support for relation schemas. See #1771

  • [FEATURE] RabbitMQ: Collect the number of RabbitMQ partitions per node. See #1715 (Thanks @ulich)

  • [FEATURE] Supervisor: Option to select processes to monitor by regex name match. See #1747 (Thanks @ckrough)

  • [FEATURE] System: Collect %guest CPU time. See #1845

  • [IMPROVEMENT] Agent Metrics: Move stats log's level to DEBUG. See #1885

  • [IMPROVEMENT] Core: Log collector runs's exceptions. See #1888

  • [IMPROVEMENT] CouchDB: Fail gracefully when one or more individual databases are not readable by the configured user. See #1760

  • [IMPROVEMENT] Docker: Add an image_repository tag to the docker check. See #1691

  • [IMPROVEMENT] Windows Event Viewer: Better configuration YAML example file. See #1734

  • [IMPROVEMENT] Windows: Add Datadog agent version to MSI description. See #1878

  • [BUGFIX] Agent Metrics: Fix the configuration YAML example file rights. See #1725

  • [BUGFIX] Amazon EC2: Update metadata endpoint list to avoid redirections. See #1750 (Thanks @dspangen)

  • [BUGFIX] Btrfs: Track usage based on used bytes instead of free bytes. See #1839 (Thanks @pbitty)

  • [BUGFIX] Couchbase: Send service check tags on OK status. See #1722 #1776

  • [BUGFIX] Docker: Fallback when Docker Remote API /events returns an invalid JSON. See #1757

  • [BUGFIX] Docker: Kubernetes support -new cgroups path-. See #1759

  • [BUGFIX] Docker: Strip newlines from API responses to avoid parsing issues. See #1727

  • [BUGFIX] Google Compute Engine: Update hostname to be unique. See #1736, #1737

  • [BUGFIX] HTTP Check: Handle requests timeout exceptions to send the appropriate service check. See #1761

  • [BUGFIX] JMXFetch: Do not override service checks's host tag with JMX host. See #66

  • [BUGFIX] JMXFetch: Do not send service check warnings on metric limit violation. See #73

  • [BUGFIX] JMXFetch: Fix collector logs being duplicated to JMXFetch ones. See #1852

  • [BUGFIX] JMXFetch: Fix indentation in the configuration YAML example file. See #1774

  • [BUGFIX] Mesos: Do not fail if no cluster name is found. See #1843

  • [BUGFIX] Mesos: Fix AttributeError on non leader nodes. See #1844

  • [BUGFIX] MongoDB: Clean password from state changed events. See #1789

  • [BUGFIX] MySQL: Close connection when complete. See #1777 (Thanks @nambrosch)

  • [BUGFIX] Network: Normalize instance_name tags to avoid mismatch and backward incompatiblity. See #1811

  • [BUGFIX] PgBouncer: Collected metrics were wrong. See #1902.

  • [BUGFIX] Process: Restore Windows support. See #1595, #1883

  • [BUGFIX] RabbitMQ: Fix ValueError error when an absolute queue name is used. See #1820

  • [BUGFIX] Redis: Handle the exception when CONFIG command is disabled. See #1755

  • [BUGFIX] Redis: Switch redis.stats.keyspace_* metrics from gauge to rate type. See #1891

  • [BUGFIX] Unix: Fix incorrect conversion of metric. See #1718, #1912

  • [BUGFIX] Windows Event Viewer: Fix indentation in the configuration YAML example file. See #1725

  • [BUGFIX] Windows: Fix developer mode configuration on Windows. See #1717

  • [BUGFIX] WMI: Fix errors when a property does not exist or has a non digit value. See #1800, #1846, #1889

  • [OTHER] Mesos: Deprecate previous generic check in favor of the Mesosphere master and slave specific checks introduced in the 5.4.0 release. See #1822

  • [OTHER] Mac OS X: Fix upgrade of the agent with DMG package. See #48

5.4.7 / 09-11-2015

Windows Only



  • [BUGFIX] Fix adodbapi dependency issue with Windows MSI Installer. See #1907

5.4.6 / 09-08-2015



  • [BUGFIX] Disk: Force CDROM (iso9660) exclusion. See #1786
  • [BUGFIX] Disk: Recalculate disk.in_use to make consistent with df's 'Use% metric'. See #1785
  • [BUGFIX] Gohai: Improve signal handling for df timeout. See #16
  • [BUGFIX] Process: Correctly handle disappearing PID. See #1721 #1772

5.4.5 / 08-20-2015

Datadog Agent container only



  • [IMPROVEMENT] Docker: Support for Docker 1.8 version. See #1831

5.4.4 / 08-05-2015

Source Install only



  • [BUGFIX] Update ntplib to 0.3.3 to fix source install requirements. See #1792

5.4.3 / 07-13-2015

Linux or Source Install only



  • [SECURITY] The deb and rpm packages now bundle OpenSSL 1.0.1p. For more details, see the security advisory.
  • [BUGFIX] Docker: Do not fail when the event API returns a bad JSON response. See #1757

5.4.2 / 06-18-2015

Linux or Source Install only



  • [BUGFIX] Disk: Strict backward compatibility with old disk check. See #1710
  • [BUGFIX] Etcd: Do not query leader endpoint on followers (was causing check failure). See #1709

5.4.1 / 06-16-2015

Linux or Source Install only



  • [BUGFIX] Disk: Get metrics only from physical disk by default. See #1700
  • [BUGFIX] Kafka: Fix indentation issue in the configuration YAML example file. See #1701

5.4.0 / 06-16-2015


New integration(s)

  • Mesosphere master
  • Mesosphere slave

Updated integrations

  • Disk
  • Docker
  • Elasticsearch
  • etcd
  • Fluentd
  • HAProxy
  • HTTP Check
  • JMXFetch
  • Kafka consumer
  • Mesosphere
  • MySQL
  • NTP
  • PostgreSQL
  • Process
  • SQL Server
  • System
  • TCP Check
  • Varnish
  • WMI

Agent Developer mode

The Agent Developer Mode allows the user to collect a wide array of metrics concerning the performance of the agent itself. It provides visibility into bottlenecks when writing an AgentCheck and when making changes to the collector core. For more information, see our wiki page.

Deprecation notice

Old agent commands

Old command line tools dd-agent, dd-forwarder and dogstatsd are deprecated as for 5.4.0. dd-forwarder & dogstatsd will be removed in a future version, and dd-agent functions will be limited to check, jmx and flare. For more information, see our wiki page See #1457, #1569

Ganglia integration

The Ganglia integration is deprecated and will be removed in a future version of the Datadog Agent. See #1621


  • [FEATURE] Agent developer mode. See #1577

  • [FEATURE] Application names to tags ("dd_check:appname") support. See #1570

  • [FEATURE] Service metadata support. See #1611

  • [FEATURE] Dogstream: Add filename globing support. See #1550 (Thanks @gtaylor)

  • [FEATURE] Elasticsearch: Add pending tasks metrics. See #1507

  • [FEATURE] Elasticsearch: Add tags to events. See #1444

  • [FEATURE] etcd: Add etcd latency metrics. See #1429

  • [FEATURE] Flare: Add commands standard error stream to content. See #1586

  • [FEATURE] Fluentd: Add type tag support. See #1623 (Thanks @yyamano)

  • [FEATURE] HAProxy: Add new time metrics available in 1.5. See #1579 (Thanks @warnerpr-cyan)

  • [FEATURE] HTTP/TCP Check: Add support for custom tags. See #1642

  • [FEATURE] JMXFetch: Add service check count to JMX checks statuses. See #1559

  • [FEATURE] Mesosphere: New checks for masters -reporting metrics from leaders- and slaves -reporting metrics from the selected tasks-. See #1535

  • [FEATURE] MySQL: Add threads running metrics. See #1613 (Thanks @polynomial)

  • [FEATURE] PHP-FPM: Add custom ping reply support. See #1582 (Thanks @squaresurf)

  • [FEATURE] System: Add system swap metrics. See #1549

  • [IMPROVEMENT] Limit process restart attempts on Windows on a specific time frame. See #1664

  • [IMPROVEMENT] Only start the Collector and Dogstatsd when needed. See #1569

  • [IMPROVEMENT] Use internal /run for temporary pid, pickle and JMXFetch files. See #1569, #1679

  • [IMPROVEMENT] Disk: New check based on psutil replaces the old system check. See #1596

  • [IMPROVEMENT] JMXFetch: Run JMXFetch as dd-agent user. See #1619

  • [IMPROVEMENT] NTP: Use Datadog NTP pool. See #1618

  • [IMPROVEMENT] Process: Cache AccessDenied failures and PID list. See #1595

  • [IMPROVEMENT] SQL Server: Set a timeout for each SQL command, default to 30s. See #1446

  • [BUGFIX] Cast service checks messages to strings. See #1617

  • [BUGFIX] Fix incorrect EC2 metadata resulting from the no proxy environment settings. See #1650 #1594

  • [BUGFIX] Uses NTP check settings to avoid failures in status checks. See #1651 #1558

  • [BUGFIX] Disk: Fix a bug where all devices were ignored if the device blacklist regex was empty. See #1666

  • [BUGFIX] Docker: Fix intermittent failures (bad support) when a container has no name. See #1565

  • [BUGFIX] Elasticsearch: Fix data being illegitimately filtered out when the local node reports under a different hostname. See #1657

  • [BUGFIX] HTTP Check: Fix status type errors in service check. See #1644

  • [BUGFIX] JMXFetch: Clean previous JMX status python file at start. See #1655

  • [BUGFIX] JMXFetch: Fix jmx agent commands false warning. See #1612.

  • [BUGFIX] JMXFetch: Format service check names prefix names to strip non alphabetic characters.

  • [BUGFIX] JMXFetch: Properly exit on Windows when a specified file is created. See #1643

  • [BUGFIX] JMXFetch: Rename 'host' bean parameter to 'bean_host' in tags to avoid conflicts.

  • [BUGFIX] JMXFetch: Support bean names that have an attribute with an empty value.

  • [BUGFIX] Kafka consumer: Add timeout for ZooKeeper and Kafka connections. See #1592 #1589

  • [BUGFIX] Mesosphere: Cast error messages to strings. See #1614 [TEST?]

  • [BUGFIX] PostgreSQL: Ignore rdsadmin database in PostgreSQL check to avoid permission error. See #1590 (Thanks @etrepum)

  • [BUGFIX] PostgreSQL: Properly log bugs for custom metrics. See #1633 (Thanks @orenmazor)

  • [BUGFIX] SQL Server: Fix collector freezes when connection is failing. See #1640

  • [BUGFIX] SQL Server: Properly close cursor, avoid leaks. See #1631

  • [BUGFIX] SQL Server: Send fractional metrics as floats. See #1616

  • [BUGFIX] Varnish: Allow tags in varnish 3 XML style parsing. See #1645

  • [BUGFIX] WMI: Cast tag values as strings. See #1630

5.3.2 / 04-29-2015

Debian only



  • [BUGFIX/FEATURE] Native support of systemd with a new service file

5.3.1 / 04-22-2015

Windows only



  • [BUGFIX] JMXFetch: Fix bootloop issue when no JMX integration is set. See #1561

5.3.0 / 04-16-2015


New integration(s)

Updated integrations

  • Cassandra
  • ElasticSearch
  • Gearman
  • HTTP Check
  • Jenkins
  • JMXFetch
  • Kafka
  • Marathon
  • Memcache
  • nginx
  • PostgreSQL
  • Redis
  • SNMP
  • Varnish
  • WMI
  • ZooKeeper


  • [FEATURE] Add a "flare" feature to contact our support team. See #1422 & #1509
  • [FEATURE] nginx: Add a metric to track dropped connections
  • [FEATURE] Redis: Add replication metrics and status. See #1350 and #1447 (Thanks @pbitty)
  • [FEATURE] Redis: Collect slow log metrics. See #1400
  • [FEATURE] WMI: Extend tagging options: tag metrics with a target class property, or a set of fixed values. See #1388
  • [FEATURE] PostgreSQL: Add support for custom queries, StatIO metrics, and database size metric. See #1395 and #1415 (Thanks @ipolishchuk and @adriandoolittle)
  • [FEATURE] Kafka: Add support of kafka >= 0.8.2. See #1438 (Thanks @patrickbcullen)
  • [FEATURE] Cassandra: Add more storage proxy metrics to default config. See #1393
  • [FEATURE] SNMP: Add support of SNMPv1. See #1408 (Thanks @slushpupie)
  • [FEATURE] Jenkins: Add support of version >= 1.598. See #1442
  • [FEATURE] JMX Checks: Add service checks
  • [FEATURE] JMX Checks: Add support of list of filters. See
  • [FEATURE] Varnish: Add support of Varnish 4.x. See #1459 and #1461
  • [FEATURE] HTTP Check: Add the possibility to test for the content of the response. See #1297, #1326 and #1390 (Thanks @chrissnel and @handigarde)
  • [IMPROVEMENT] JMXFetch: Move JMXFetch to its own entry in the supervisor
  • [IMPROVEMENT] Switch http library used in checks to requests. See #1399
  • [IMPROVEMENT] NTP Check: Enable by default
  • [IMPROVEMENT] EC2 tags: require only DescribeTags permission. See #1503 (Thanks @oremj)
  • [BUGFIX] JMXFetch: Add default min and max heap size
  • [BUGFIX] PostgreSQL: Fix "Metric has an interval of 0 bug". See #1211 and #1396
  • [BUGFIX] Marathon: Fix bad url construction. See #1278 and #1401
  • [BUGFIX] Zookeeper: Fix misleading metric name. See #1443 and #1383
  • [BUGFIX] Proxy settings: Cast proxy port to an integer. See #1414 and #1416
  • [BUGFIX] Support EC2 tag discovery in all regions. See #1332
  • [BUGFIX] Source installation: Fix "error: no such option: --use_simple_http_client". See #1454
  • [BUGFIX] Memcache: Fix bad support of multi instances. See #1490
  • [BUGFIX] Gearman: Fix bad support of multi instances. See #1476
  • [BUGFIX] HTTP Check: Fix for servers using SNI
  • [BUGFIX] ElasticSearch: Fix bad support of multi instances. See #1487
  • [BUGFIX] Core: Do not use proxy for local connection. See #1518

5.2.3 / 03-30-2015

Windows only



  • [BUGFIX] Fix vSphere service check

5.2.2 / 03-20-2015

Linux or Source Install only



  • [SECURITY] The deb and rpm packages now bundle OpenSSL 1.0.1m
  • [BUGFIX] Fix "pidfile /tmp/ already exists" bug. See #1435
  • [BUGFIX] Fix bundling of rrdtool python binding

5.2.1 / 02-20-2015

Linux or Source Install only



  • [BUGFIX] varnish: fix regression, bad argument in _parse_varnishstat. See #1377 (Thanks @mms-gianni)
  • [BUGFIX] source install: move pysnmp and pysnmp-mibs to optional reqs. See #1380
  • [BUGFIX] etcd: service check OK is now returned. See #1379
  • [BUGFIX] varnish: fix varnishadm sudoed call with subprocess. See #1389

5.2.0 / 02-17-2015


New and updated integrations

  • CouchDB
  • Couchbase
  • Docker
  • ElasticSearch
  • etcd
  • fluentd
  • Gearman
  • GUnicorn
  • HTTPCheck
  • JMXFetch
  • KyotoTycoon
  • Marathon
  • Mesos
  • Network
  • Postgresql
  • Process
  • Riak
  • RiakCS
  • SNMP
  • Supervisor
  • TeamCity
  • TokuMX
  • Varnish
  • VSphere
  • Windows Event Viewer
  • Windows Services
  • Windows System metrics


Starting from this version of the agent, the default endpoint URL is replaced by an ad-hoc by version endpoint: We might use other endpoints to better route the traffic on our end in the future. See more details at


  • [FEATURE] Dogstatsd: Add an option to namespace all metrics. See #1210 (Thanks @igor47)

  • [FEATURE] Couchdb: Add a service check. See #1201

  • [FEATURE] Couchbase: Add a service check. See #1200

  • [FEATURE] Gearman: Add a service check. See #1203

  • [FEATURE] GUnicorn: Add a service check. See #1163

  • [FEATURE] KyotoTycoon: Add a service check. See #1202

  • [FEATURE] Marathon: Add a service check. See #1205

  • [FEATURE] Mesos: Add a service check. See #1205

  • [FEATURE] Riak: Add a service check. See #1187

  • [FEATURE] SNMP: Add a service check. See #1236

  • [FEATURE] TokuMX: Add a service check. See #1173

  • [FEATURE] Varnish: Add a service check. See #1213

  • [FEATURE] VSphere: Add a service check. See #1238

  • [FEATURE] VSphere: Allow host filtering. See #1226

  • [FEATURE] HTTPCheck: Check for SSL certificate expiration. See #1152

  • [FEATURE] etcd: Add new etcd integration. See #1235 (Thanks @gphat)

  • [FEATURE] Process: Better SmartOS support. See #1073 (Thanks @djensen47)

  • [FEATURE] Windows Event Viewer: Allow filtering by id. See #1255

  • [FEATURE] Windows Services: Monitor state of Windows Services. See #1225

  • [FEATURE] Windows: Get more system metrics regarding memory and disk usage.

  • [FEATURE] Windows: Better GUI

  • [FEATURE] Adding “min” metric to histograms. See #1219

  • [FEATURE] Activemq: New ActiveMQ XML check that collect more metrics. See #1227 (Thanks @brettlangdon)

  • [FEATURE] TeamCity: New TeamCity integration. See #1171 (Thanks @AirbornePorcine)

  • [FEATURE] RiakCS: Add a RiakCS Integration. See #1101 (Thanks @glickbot)

  • [FEATURE] FluentD: Add a FluentD integration. See #1080 (Thanks @takus)

  • [FEATURE] Docker: Configurable image count collection. See #1345

  • [FEATURE] SNMP: Integer and Integer32 metric types support. See #1318

  • [FEATURE] JMXFetch: Fetch more JVM (Non)Heap variables by default. See #42

  • [BUGFIX] Docker: Filter events too. See #1285

  • [BUGFIX] ElasticSearch: Handle Timeout. See #1267

  • [BUGFIX] ElasticSearch: Only query the local node. See #1181 (Thanks @jonaf)

  • [BUGFIX] Marathon: Fix check on Marathon >= 0.7. See #1240

  • [BUGFIX] Network: Fix interface skipping. See #1260 (Thanks @sirlantis)

  • [BUGFIX] Postgreql: Fix service check. See #1273

  • [BUGFIX] Postgresql: Fix BGW metrics. See #1272 (Thanks @ipolishchuk)

  • [BUGFIX] Postgresql: Fix buffers_backend_fsync. See #1275

  • [BUGFIX] SNMP: Fix "tooBig" SNMP error. See #1155 (Thanks @bpuzon)

  • [BUGFIX] Zookeeper: Fix bad command sending.

  • [BUGFIX] ElasticSearch: Fix host tagging. See #1282

  • [BUGFIX] SNMP: Fix non-increasing OID issue. See #1281

  • [BUGFIX] Dogstatsd: Properly handle UTF-8 packets. See #1279

  • [BUGFIX] SQLServer: Fix for Latin1_General_BIN Collection Servers. See #1214 (Thanks @PedroMiguelFigueiredo)

  • [BUGFIX] FreeBSD: Get full interface name. See #1141 (Thanks @mutemule)

  • [BUGFIX] SNMP: Fix a 'Missing OID' issue. See #1318

  • [BUGFIX] JMXFetch: Fix a memory leak issue. See #30

  • [BUGFIX] Windows Event Log: Fix a timezone issue. See #1370

5.1.1 / 12-09-2014


Updated integrations

  • MongoDB


  • [BUGFIX] MongoDB: Fix TypeError that was happening in some cases. See #1222
  • [BUGFIX] BTRFS: Handle "unknown" usage type. See #1221
  • [BUGFIX] Windows: When uninstalling the Agent, the uninstaller was mistakenly telling the user that the machine would reboot. This is fixed.

5.1.0 / 11-24-2014



  • Pup is now removed from the Datadog Agent
  • The "ALL" parameter in the process check is deprecated and will be removed in a future version of the agent.
  • The Windows installer does not require the .NET framework anymore.

New and updated integrations

  • PostgreSQL
  • Directory
  • Jenkins
  • MongoDB
  • Process
  • ElasticSearch
  • IIS
  • ZooKeeper
  • Memcached
  • SSH
  • System Core


  • [FEATURE] Add Service Checks for the following integration:

    • Apache
    • HAProxy
    • Lighttpd
    • NginX
    • NTP
    • HTTP
    • TCP
    • Process
    • ElasticSearch
    • IIS
    • ZooKeeper
    • Memcached
    • MongoDB
    • SQL Server
    • MySQL
    • PostgeSQL
  • [FEATURE] PostgreSQL: Pick up per-table size stats. See #1105

  • [FEATURE] PostgreSQL: Collect locks per mode and bgwriter metrics. See #1019

  • [FEATURE] Directory Check: Let the possibilty to tag metrics at the file level. See #1041 (Thanks @Osterjour)

  • [FEATURE] Jenkins: Add result and build number to events tags. See #1068 (Thanks @jzoldak)

  • [FEATURE] Add a SSH Check. See #1117

  • [FEATURE] Add a check to collect metrics from BTRFS. See #1123

  • [FEATURE] Add a check to collect system core metrics. See #1124

  • [FEATURE] DogStatsD recognizes and uses host and device tags as metric attributes. See #1164.

  • [FEATURE] Docker: Revamp events and add more options. See #1162.

  • [FEATURE] Docker: Collect relevant Docker metrics by default, make the others optional. See #1207.

  • [FEATURE] Docker: Improve Docker metrics tagging. See #1208 and #1218.

  • [BUGFIX] Jenkins: Fix when build does not yet have results. See #1060 (Thanks @jzoldak)

  • [BUGFIX] PostgreSQL: If connection drops, re-establish at next run. See #1105

  • [BUGFIX] MongoDB: Add logging of serverStatus errors. See #1065 (Thanks @igroenewold)

  • [BUGFIX] Docker: Fix various time-outs and errors. See #1162.

5.0.5 (Every platform) / 10-31-2014


This release fixes a bug on servers that are configured in local time instead of UTC Time. If your server's clock is configured to use daylight saving time, your server might stop sending metrics for up to one hour when the Daylight Saving Time ends or until the Agent is restarted after the Daylight Saving Time ends.

We highly recommend to upgrade to this version if your server is configured in local time.

5.0.4 (deb package, rpm package) / 10-17-2014


This is a security update regarding POODLE (CVE-2014-3566).

The Omnibus package will now bundle OpenSSL 1.0.1j without support of SSLv3 (no-ssl3 flag) and Python 2.7.8 with a patch that disables SSLv3 unless explicity asked

This Omnibus package also adds support of the sqlite3 library for Python.

5.0.3 (Windows only)


vSphere check:

  • [FEATURE] Batching jobs to cache the infrastructure of vCenter when autodiscovering Hosts/VMs is configurable
  • [BUGFIX] Fix ESXi host tags not being correctly set
  • [BUGFIX] Fix metadata reset so that metrics processing is not stopped when refreshing metadata
  • [BUGFIX] Fix thread pool crash when one thread would not terminate gracefully

5.0.2 (Windows only)


vSphere check:

  • [FEATURE] Changed the event filter to remove login events by default
  • [BUGFIX] Duplicate tags on VMs and host
  • [BUGFIX] Ignore duplicate events about VM migrations

5.0.1 (Windows only)


[FEATURE] Releasing the vSphere check. This is a new integration able to fetch metrics and events from vCenter.

5.0.0 / 08-22-2014



This is a major version of the Datadog-Agent.

  • On Linux: Packaging of the Agent has changed for RPM and DEB packages.

  • On Windows: This release has multiple fixes, see the list below. Warning: The way CPU metrics are collected has changed and will be more accurate, you might see some changes in the graphs.

What will break ?

  • MySQL integration: If you see this error: OperationalError(2003, 'Can\'t connect to MySQL server on \'localhost\' ((1045, u"Access denied for user \'datadog\'@\'\'...) the Datadog user will need to be modified from 'datadog'@'localhost' to 'datadog'@'' (your host IP). You can do this by running:

         $ mysql -p mysql
         # UPDATE user SET Host = '' WHERE User = 'datadog';
  • If you were using a custom check that needed python dependencies you will have to reinstall them using the bundled pip:


sudo /opt/datadog-agent/embedded/bin/pip install YOUR_DEPENDENCY ```

  • Configuring checks in datadog.conf for checks.d is deprecated and won't work anymore. Please configure your checks by editing the yaml files in the conf.d directory.

How to upgrade?

See this Wiki page

New and updated integrations

  • Docker
  • ElasticSearch
  • Kafka
  • Kafka consumer
  • NTP
  • HDFS
  • Postgres
  • Process
  • Redis
  • SNMP
  • SQL Server


  • [FEATURE] Add support of Centos 7 and Fedora Core 19-20
  • [FEATURE] Add a NTP check. See #971
  • [FEATURE] Add an option to instrument check runs time. See #1013
  • [FEATURE] Add derived Redis metrics. See #1015
  • [FEATURE] Add an SNMP Check. See #299
  • [FEATURE] Redis: Adds support for checking length of more types. See #996
  • [FEATURE] Let the possibility to exclude some disks for Windows disks check. See #1008
  • [FEATURE] Collect more Docker metrics. See #1027
  • [FEATURE] Docker check: work inside a Docker container. CoreOS support. See #1001
  • [FEATURE] HDFS: Add support for HA mode. See #1018. Warning: It requires snakebite >= 2.2.0
  • [BUGFIX] Support Windows EOL \r character. See #1023
  • [BUGFIX] Fix the collection of cpu metrics (>100%) on Windows. See #653
  • [BUGFIX] Fix connection error on Windows 2008 SP2. See #1014
  • [BUGFIX] Dogstreams on windows: Allow to specify custom parser. See #887
  • [BUGFIX] ElasticSearch: Fix elasticsearch metrics according to different ES versions: See #1024
  • [BUGFIX] Process check: Fix check on some version of psutil. See #958
  • [BUGFIX] Fix init script on Centos/RHEL when dogstatsd is disabled. See #1002
  • [BUGFIX] Fix kafka metrics by sending them as gauges instead of rate. See #1029
  • [BUGFIX] Kafka consumer: Support version 0.9 of python-kafka. See #1028
  • [BUGFIX] Postgres: fix columns to retrieve when two different postgres version on the same host. See #1035
  • [BUGFIX] Fix multiple Docker check bugs. See #1017 #1031
  • [BUGFIX] Lets the possibility to override curl behavior when using a proxy that returns a 302. See #1036
  • [BUGFIX] SQL Server: Detect automatically counter types to report them correctly. See #1069
  • [BUGFIX] Report Docker memory page metrics as rates.

4.4.0 / 06-24-2014


Integrations affected

  • Docker
  • Redis
  • Memcached
  • MySQL
  • PostgreSQL


  • [BUGFIX] Docker: Don't raise Exception if we fail to get some Docker croup info. See #981
  • [BUGFIX] Docker: Don't raise Exception if no Docker containers are running. See #980
  • [BUGFIX] Docker: Fix integration timeout issue. See #963
  • [ENHANCEMENT] Let the possibility to disable metadata collection from See #975
  • [FEATURE] Redis: Add a metric to track key length. See #962
  • [FEATURE] MySQL: Collect more metrics. See #972
  • [BUGFIX] MySQL: Only collect metrics from /proc on linux machines. See #984
  • [BUGFIX] PostgreSQL: Handle negative replication delay. See #977
  • [ENHANCEMENT] Collect more Memcached stats. See #982
  • [BUGFIX] Remove Content-Length header in CONNECT HTTP Requests (when using a proxy), as CONNECT Requests shouldn't have that header and some proxies don't support it.

4.3.1 / 06-03-2014


Linux or Source Install only

Integrations affected

  • Docker
  • HAProxy


  • [IMPROVEMENT] Don't collect Docker total_ metrics by default. See #964
  • [BUGFIX] Report Docker CPU metrics as rate. See #964
  • [BUGFIX] Add HAProxy reporter name in HAProxy event's titles. See #960

4.3.0 / 05-22-2014


Integrations affected

  • MongoDB
  • Process
  • CouchDB
  • Docker
  • HAProxy
  • Marathon
  • Memcached
  • Mesos
  • MySQL
  • TokuMX
  • ElasticSearch
  • Network


  • [BUGFIX] Fix incorrect open file descriptors metric name in process check: See #904. Warning: As the metric name will be changed. Your dashboards will have to be updated.
  • [FEATURE] Add some mongo2.2+ metrics to check: See #951 (Thanks to @ckrough)
  • [FEATURE] Collect checks statuses: See #922
  • [FEATURE] CouchDB: Support basic authentication: See #930
  • [FEATURE] Docker: Support Docker 0.11
  • [FEATURE] Docker: Collect events from the events api
  • [FEATURE] HAProxy: Allow collection of metrics per host with the option collect_status_metrics_by_host: See #935
  • [BUGFIX] HAProxy: Fix inaccuracy of count_per_status metric: See #940
  • [BUGFIX] HAProxy: Update event's titles: See #935
  • [FEATURE] Add Marathon integration: See #921 (Thanks to @graemej)
  • [FEATURE] Add Mesos integration: See #919 (Thanks to @graemej)
  • [FEATURE] Memcached: Add delete_misses/delete_hits: See #928 (Thanks to @arthurnn)
  • [BUGFIX] MySQL: Only collect MySQL metrics from /proc on unix machines: See #947
  • [BUGFIX] MySQL: Fix duplicate metric bug: See #899
  • [BUGFIX] Varnish: Fix a bug that was causing tags to be continuously added in some cases.
  • [FEATURE] Add an option to disable Dogstastsd: See #927
  • [FEATURE] Memcached: Add support for local unix socket connections: See #891 (Thanks to @mike-lerch)
  • [FEATURE] Add TokuMX integration: See #933 (Thanks to @leifwalsh)
  • [BUGFIX] ElasticSearch: Added correct GC metrics for ES 1.x: See #900 (Thanks to @joningle)
  • [FEATURE] Network: Add additional metrics for TCP: See #949

4.2.2 / 04-25-2014


Windows Only

Integrations affected

  • Redis


  • [FEATURE] Support Redis check on Windows: See #917

4.2.1 / 04-09-2014


Integrations affected

  • ElasticSearch
  • Nginx
  • Process
  • Postgres


  • [BUGFIX] Fix bug in Dogstatsd in the bucketing change: See #894
  • [BUGFIX] Revert to the Simple HTTP Client by default in the forwarder as it's causing 599 in some cases. It's now configurable in datadog.conf
  • [FEATURE] Support for OpenShift cartridges: See #875
  • [PERFORMANCE] Compress Dogstatsd payloads: See #893
  • [BUGFIX] Fix process check compatibility with psutil 2.0: See #863
  • [FEATURE] Support additional NGINX Plus metrics: See #876
  • [PERFORMANCE] Better handling of external clusters in Elasticsearch check: See #883
  • [BUGFIX] Fix an issue that is causing a high number of tags for postgresql metrics when custom tags are enabled.

4.2.0 / 03-25-2014


Integrations affected

  • Couchbase
  • Docker
  • ElasticSearch
  • HAProxy
  • Kafka consumer
  • Kafka server
  • MongoDB
  • MySQL
  • PostgreSQL
  • Process
  • Google Compute Engine
  • JMX Checks: Cassandra, Tomcat, Solr, ActiveMq, JMX, Kafka


  • [PERFORMANCE] Disable pup by default
  • [ENHANCEMENT] Use JMXFetch 0.3.0 ( Changelog )
  • [ENHANCEMENT] Metric limit for JMX Checks is raised to 350 metrics per instance
  • [FEATURE] Add a "configtest" command alias of "configcheck": See #838
  • [FEATURE] Add a Docker integration: See #844 (Thanks to @steeve)
  • [FEATURE] ElasticSearch: Support for newer versions (>= 0.90.10)
  • [FEATURE] ElasticSearch: Add a metric to monitor cluster status: See #827 (Thanks to @igor47)
  • [FEATURE] HAProxy: Add availability metrics: See #834 (Thanks to @igor47)
  • [FEATURE] Add a Kafka consumer check (Requires python 2.6 or python 2.7): See #810
  • [FEATURE] Add a Kafka server check: See #810
  • [FEATURE] MongoDB: Support SSL connections to server: See #848
  • [FEATURE] Collect tags and hostname for Google Compute Engine hosts
  • [FEATURE] PostgreSQL: Support check on Windows
  • [ENHANCEMENT] Align the reporting of StatsD metrics to fixed time intervals for consistency in aggregation
  • [PERFORMANCE] StatsD events are now batched: See #852
  • [PERFORMANCE] Add an optional timeout parameter to Couchbase integration: See #826
  • [PERFORMANCE] Use Tornado CurlAsyncHTTPClient by default over the SimpleHTTPClient
  • [BUGFIX] MySQL: Fixed warning on SHOW SLAVE STATUS: See #809 (Thanks to @ive)
  • [BUGFIX] PostgreSQL: Reset the connection if it seems broken: See #784
  • [BUGFIX] Process: Do not fail on older Linux Kernels: See #849
  • [BUGFIX] Windows: Do not restart pup on windows if it's not enabled: See #815
  • [BUGFIX] JMX Checks: Properly ensure that only one instance of JMXFetch is running

4.1.0 / 2014-02-04


Integrations affected

  • PostgreSQL
  • ElasticSearch
  • Lighttpd
  • Nginx
  • HAProxy
  • MongoDB
  • Redis
  • Varnish
  • Couchbase


  • [FEATURE] Support for older versions of psycopg2
  • [FEATURE] New tool to help configuring JMX Checks: See
  • [FEATURE] Add basic authentication to Couchbase check #787
  • [FEATURE] Add basic authentication to ElasticSearch check #806
  • [FEATURE] Add basic authentication to Lighttpd check
  • [FEATURE] Add basic authentication to Nginx check
  • [FEATURE] Calculate used session percentage in HAProxy check #752 (Thanks to @walkeran)
  • [FEATURE] Add an HAProxy metric that counts the number of active backends #729
  • [FEATURE] Turn SSL validation in http check into an option #770
  • [FEATURE] Include tags in http service check #780
  • [FEATURE] Add more metrics to MongoDB check for newer versions of MongoDB #735
  • [FEATURE] Support multiple mongo instances in MongoDB check 08be06f4c
  • [FEATURE] Add more metrics to MySQL check #72 (Thanks to @skingry and @ronaldbradford)
  • [FEATURE] Add per-table metrics to Postgresql check #760
  • [FEATURE] Add more metrics in process check with newer version of psutil
  • [FEATURE] Allow configuration of the Redis check using a unix socket path #730
  • [FEATURE] Allow multiple instances of Varnish check #490
  • [FEATURE] Add the ability to add tags to an elasticsearch instance #790 (Thanks to @clly)
  • [BUGFIX] Fix automatic start of the Agent on Windows acf368c
  • [BUGFIX] Fix Gunicorn check issue where it was failing to identify the process in some cases #706
  • [BUGFIX] Don’t fail on archived builds in Jenkins check #766 (Thanks @imlucas)

4.0.2 / 2014-01-08

Windows Only



  • [BUGFIX] Fix WMI Check

4.0.1 / 2013-12-17

Linux or Source Install only



  • [BUGFIX] Fix Postgresql check that was sending bad values in some cases.
  • [BUGFIX] Fix replication lag calculation in MySql check.

4.0.0 / 2013-12-16

This is a major version. See platform-specific pages for detailed changes.



3.10.1 / 2013-11-06

Linux or Source Install only



  • [BUGFIX] Fix Mongo Integration for newer versions of MongoDB #677
  • [BUGFIX] Fix memory metrics for Mac Os X Mavericks
  • [BUGFIX] Fix tagging issues for HTTP Check 8ab75
  • [BUGFIX] Fix local issues 4230

3.11.0 / 2013-10-08

Windows Only


Integrations Affected

  • Cassandra
  • Tomcat
  • ActiveMQ
  • SolR
  • Java
  • MySQL
  • Riak


  • [FEATURE] Make Cassandra, Tomcat, ActiveMQ, SolR, Java and MySQL integrations work on Windows
  • [FEATURE] Make pup work on Windows
  • [FEATURE] Add an additional metric to the Nginx integration #665 (Thanks to @dcrosta)
  • [FEATURE] Add additional metrics to Riak metrics #643 (Thanks to @stefan-mees)
  • [BUGFIX] Fix Service checks on Windows #632

3.9.3 / 2013-09-11

Windows Only


Integrations Affected

  • SQL Server


  • [FEATURE] Allow optional custom tags in SQL Server check (#654)
  • [BUGFIX] Fix log file location on Windows XP

3.10.0 / 2013-09-06

Linux or Source Install only


Integrations Affected

  • HTTP Check
  • Mongo
  • MySQL
  • Network
  • Process


  • [FEATURE] GUnicorn check #619
  • [FEATURE] Dogstatsd Autorestart #624
  • [FEATURE] Add tags to metrics collected by the HTTP Check #647 (Thanks to @ordenull)
  • [FEATURE] Allow MySQL check configuration via a MySQL config file #590 (Thanks to @micktwomey)
  • [FEATURE] Filter disk, io & network metrics by device #615
  • [FEATURE] Collect metrics from the MongoDB database selected in the connection string #657
  • [FEATURE] Add CPU and thread utilisation metrics to the Process check #646 (Thanks to @morskoyzmey)
  • [BUGFIX] Add a timeout to the Mongo connection #627

3.9.2 / 2013-08-29

Windows Only


Integrations Affected

  • SQL Server


  • [FEATURE] Default SQL Server to integrated security if no username/password is provided (#622)(Thanks to @jamescrowley)
  • [FEATURE] Allow skipping ssl certificate validation (useful when the agent runs behind haproxy) (#641)
  • [BUGFIX] Fix proxy support on Windows
  • [BUGFIX] Better management of config files with the GUI

3.9.1 / 2013-08-19

Windows Only


Integrations Affected

  • SQL Server
  • IIS


  • [FEATURE] Add a Management GUI to the Windows Agent for service and check management
  • [FEATURE] Log to a log file (located in C:\ProgramData\Datadog\logs )
  • [FEATURE] Create shortcuts in the Start Menu
  • [BUGFIX] Fix status page
  • [BUGFIX] Fix logging in the event viewer and only logs errors (#496)
  • [BUGFIX] Add debug info in the sql server check (#608)
  • [BUGFIX]IIS: By default use _Total, but allow a configurable list of sites (6885a97bc)

3.9.0 / 2013-08-05


Integrations Affected

  • HDFS
  • Postgres
  • MySQL
  • Jenkins
  • Nginx
  • RedisDB


  • [FEATURE] New HDFS check added (#551) (thanks to @dcrosta)
  • [FEATURE] New Directory check added (#581) (thanks to @brettlangdon)
  • [FEATURE] Events can now be sent to the Agent DogStatsD server from supported client libraries (#532)
  • [FEATURE] HTTP check now supports custom headers (#541) (thanks to @tomduckering)
  • [FEATURE] Optional response_time metric has been added to TCP and HTTP checks
  • [FEATURE] info command will exit with a non-zero value when errors were displayed by the command
  • [FEATURE] Basic Jenkins metrics are now collected by Jenkins check (#567)
  • [FEATURE] A non-default port can now be specified in the MySQL check (#575)
  • [FEATURE] Logs now follow The BSD syslog Protocol (#577)
  • [BUGFIX] Expat XML Parser dependency is now installed by SmartOS Agent installation script (#450)
  • [BUGFIX] Fix collection of Postgres rollbacks metric
  • [BUGFIX] Fix Postgres integration crashing when tags are None
  • [BUGFIX] Fix version detection in MySQL check (#558)
  • [BUGFIX] MySQL InnoDB metrics are now only collected with InnoDB is enabled (#566)
  • [BUGFIX] The source status and info commands will no longer attempt to start the Agent (#512)
  • [BUGFIX] Upon a failed EC2 metadata lookup, the last successfully collected metadata will now be report (#554)
  • [BUGFIX] Nginx check no longer asserts number of connections (#569)
  • [BUGFIX] Deb and RPM start command will now poll the Agent status when starting instead of waiting a fixed amount of time (#582)
  • [BUGFIX] RedisDB check will now cast a parsed port to an integer (#600)
  • [BUGFIX] supervisord location is no longer hardcoded on Debian (#580) (Thanks to @mastrolinux)

3.8.0 / 2013-06-19



  • [FEATURE] Add status command to Debian
  • [FEATURE] Debian version now uses its own supervisor config instead of using the system config
  • [FEATURE] Add -v option to info command, which currently gives stack traces for errors that occurred during checks
  • [FEATURE] Add I/O metrics to OS X (#131)
  • [BUGFIX] Log exception when dogstatsd server fails to start (#480)
  • [BUGFIX] Fix Error: Invalid user name dd-agent appearing during source install (#521)
  • [BUGFIX] Debian and Red Hat init.d scripts now verify that /etc/dd-agent/datadog.conf is present before launching supervisor(#544)
  • [BUGFIX] Fix AttributeErrors for timeout_event and status_code_event in Riak check (#546)

3.7.2 / 2013-05-22



  • [FEATURE] Fix redis check when used with additional tags (#515)

3.7.1 / 2013-05-21



  • [FEATURE] Add basic auth support for apache check (#410)
  • [FEATURE] Support any redis parameter during the connection (#276)
  • [FEATURE] Better launching script for source install
  • [BUGFIX] Fix process check (Missing import and support version 0.4 of psutil) (#502)
  • [BUGFIX] Fix JVM Heap issue when launching java process ( Disable memory consumption watcher by default) (#507)
  • [BUGFIX] Info page shows errors when failing to initialize a check.d (#427)
  • [BUGFIX] Added file option to supervisorctl stop arg too (#498) (Thanks to @mastrolinux)
  • [BUGFIX] Fix mysql version detection (#501)

3.7.0 / 2013-05-08



  • [FEATURE] Restart the agent if it uses too much memory (#426) (Thanks to @echohead)
  • [FEATURE] Port Memcache to checks.d (#390)
  • [FEATURE] Add a process memory check (#434) (Thanks to @mastrolinux)
  • [FEATURE] Add a gearman check (#435) (Thanks to @CaptTofu)
  • [FEATURE] Add a Web Info Page to check the status of the agent (http://localhost:17125/status) (#483)
  • [FEATURE] Create an info page for the source install (#481)
  • [FEATURE] Add a “warning” method to the AgentCheck class that will display warnings in the info page
  • [BUGFIX] Customizable Java directory for JMX Checks (#472)
  • [BUGFIX] Do not try to write logs in /var/log when using the source install (#478)
  • [BUGFIX] Use a Unix socket in supervisor for the source installation
  • [BUGFIX] Display more information when the agent stops because there is no valid hostname (#475)

3.6.4 / 2013-04-25

Windows only


Bug fixes

  • IIS: Use Total metrics and calculate rates in the Agent instead of using PerSec metrics. (#465)
  • IIS: Optionally give a list of sites to pull metrics from, defaulting to all.

3.6.3 / 2013-04-14



  • [BUGFIX} Customizable field names for cacti check (#404)
  • [BUGFIX} Enable replication monitoring by default for old style check configuration for mysql
  • [BUGFIX} Always collect metrics for config specified queues/nodes for rabbitmq

3.6.2 / 2013-04-05



  • [FEATURE] Port MySQL to checks.d (#408) (Thanks to @CaptTofu)
  • [FEATURE] Add KyotoTycoon Check (#426) (Thanks to @dcrosta)
  • [FEATURE] Add command line option to run a particular agent check (#408)
  • [BUGFIX} Force include elementtree.ElementTree in Windows install (#423)
  • [BUGFIX} Fix elasticsearch check for version < 0.19 (#419)
  • [BUGFIX} Disable HAProxy events by default
  • [BUGFIX} Aggregate RabbitMq Metrics over queues and nodes
  • [BUGFIX} Better hostname detection
  • [BUGFIX} Fix broken json serialization in centos5 (#422)

3.6.1 / 2013-03-21



  • [FEATURE] Port Jenkins to checks.d
  • [FEATURE] Lighttpd check now supports Lighttpd 2.0 (#412) (Thanks to @brettlangdon)
  • [FEATURE]Additional configurable checks.d directory (#413) (Thanks to @brettlangdon)
  • [BUGFIX] Better Jenkins check performance (reduce CPU consumption) (#402)
  • [BUGFIX] Fix Graphite listener (#415)
  • [BUGFIX] Less logging for pup (#414)

3.6.0 / 2013-03-13



  • [FEATURE] The agent can now run behind a web proxy (#377)
  • [FEATURE] MongoDB now supports multiple instances running on the same host (#397)
  • [FEATURE] Additional network metrics (#396)
  • [FEATURE] lighttpd check (#385)
  • [FEATURE] Allow pup to bind to a specific interface (#394). Thanks to @shamada-kuuluu
  • [FEATURE] Add a partial response in HTTP Check (#375). Thanks to @dcrosta
  • [FEATURE] WMI checks support advanced configuration (#359)
  • [FEATURE] More reliable and consistent hostname detection (84e715c)
  • [BUGFIX] Better retry handling for JMX checks (#369)
  • [BUGFIX] Fix JMX Python 2.4 support (#401)


This is a RedHat-only release.



  • [BUGFIX] Fix dogstatsd crash on RedHat 5.x and its derivatives (#381)




  • [FEATURE] Logging overhaul: Consistent locations in /var/log/datadog/, defaults to INFO level (#297)
  • [FEATURE] Add more memcached metrics (#283). Thanks to @jkoppe
  • [FEATURE] RabbitMQ integration (#330). Thanks to @brettlangdon
  • [FEATURE] Riak integration (#332). Thanks to @stefan-mees
  • [FEATURE] Allow source file and line in Cassandra system.log. (#307)
  • [FEATURE] Port CouchDB and ElasticSearch to checks.d (#311)
  • [FEATURE] New System Metrics: system.mem.pct_usable and system.swap.pct_free (#334)
  • [FEATURE] SmartOS support (see the agent setup page)
  • [FEATURE] Invoke custom emitters from the forwarder, instead of the agent, to capture statsd output (#271. Thanks to @charles-dyfis-net)
  • [FEATURE] Allow strings to be elements in dogstatsd sets (#300)
  • [BUGFIX] Limit the number of threads used by service checks (#351)
  • [PERFORMANCE] Better JMX performance (#313, #348)
  • [BUGFIX] Fix names of some Apache metrics (#326)




  • [BUGFIX] Fix memory leaks in redis check, and potentially in custom checks.d checks that supply duplicate tags (#325)
  • [BUGFIX] Fix mongo auth issue (#318)
  • [BUGFIX] Add configurable socket timeout to zookeeper check (#310)




  • [BUGFIX] Fix memory leaks in memcache check (#278)
  • [BUGFIX] Fix umask issue (#293)
  • [BUGFIX] Fix bad error message in CentOS 5 installation (#320)


If you're having issues upgrading, please read the upgrade notes



  • [FEATURE] Check multiple Cassandra, Tomcat and Solr instances per host
  • [FEATURE] Added a JMXCheck base class which can be used to easily track metrics from services that support JMX.
  • [BUGFIX] Create /etc/dd-agent/conf.d on install
  • [BUGFIX] Reduce verbosity of the logs




  • [FEATURE] Added an info command (sudo /etc/init.d/datadog-agent info) which prints status info about the agent processes.
  • [FEATURE] Added a check for Zookeeper.
  • [BUGFIX] Fixes packaging bugs introduced in 3.4.0.
  • [BUGFIX] Agents installed with RPM will restart on upgrade (starting with the next version).
  • [BUGFIX] Fixed normalized counter rounding error.
  • [BUGFIX] By default, don't open ports other than localhost.

3.4.0 / 2012-11-28



  • [FEATURE] Added FreeBSD support, thanks to @loris.
  • [FEATURE] Removed datadog-agent and datadog-agent-base dependencies. Now you only install one package per machine (instructions are the same).
  • [FEATURE] The agent now compresses payloads sent over the wire.
  • [FEATURE] Allow custom PYTHONPATH in checks.d config (#227)
  • [FEATURE] Added new Redis metrics.
  • [FEATURE] Added normalized load, that is load per cpu.
  • [FEATURE] Port Apache, NginX and Varnish checks to checks.d.
  • [BUGFIX] #290, #291 - disable non-local traffic by default, suppress stack traces in 404s
  • [BUGFIX] #257 - More useful Apache rates not averaged from the beginning of time.
  • [BUGFIX] #277 - Run dogstatsd on older debian boxes.
  • [BUGFIX] #245 - Expire counter values.
  • [BUGFIX] #261 - Fix Windows checks.d location lookup.
  • [BUGFIX] #253 - Sum Dogstream counters.
  • [PERFORMANCE] Improved dogstatsd performance.
  • [ENHANCEMENT] Stylistic code improvements.

3.3.0 / 2012-10-25


New Features


  • [FEATURE] Added HTTP and TCP Service Checks (read the docs)
  • [FEATURE] Added the Windows Event Log Integration
  • [PERFORMANCE] Use the much faster simplejson library, if it's available, otherwise use the standard json library.
  • [BUGFIX] Fixed disk space metrics bug
  • [BUGFIX] Run dogstatsd on older OS's as well.
  • [BUGFIX] Fixed host aliasing issue.
  • [BUGFIX] Use a better query to get the Cacti device name.
  • [BUGFIX] Ensure pup uses the same JSON parsing library as the rest of the application.

3.2.3 / 2012-10-15



  • [FEATURE] Windows support is officially added.
  • [FEATURE] Added a SQL Server check.
  • [FEATURE] Added an IIS check.
  • [FEATURE] Track request_rate in HAProxy.
  • [FEATURE] Move DogstatsD to the datadog-agent-base package.

3.2.2 / 2012-10-05



  • [BUGFIX] Fixes an issue with events in checks.d where only events from the last instance would be sent.

3.2.1 / 2012-10-05



  • [BUGFIX] Fixes an issue with duplicate events being created in checks.d checks.

3.2.0 / 2012-10-05



  • [FEATURE] Add new AgentCheck interface for all future checks.
  • [FEATURE] Split checks and configuration with checks.d/conf.d.

3.1.7 / 2012-09-28



  • [BUGFIX] Fixes the case where you have the python-redis module and the check will run with default host/port even if you don't have any redis configuration. Fixes case #200.

3.1.6 / 2012-09-27



  • [BUGFIX] Fixes memcached integration bug running under Python 2.4 #201
  • [BUGFIX] Removes token from the Cassandra Stats, because it is not always a number. Fixes case #202

3.1.5 / 2012-09-21



  • [BUGFIX] Fixes network traffic reporting bug introduced in 3.1.4. If you're running 3.1.4 we recommended that you upgrade.

3.1.4 / 2012-09-21



  • [FEATURE] memcached and nginx checks now support multiple instances per host.
  • [FEATURE] Statsd: Added sets metric type. Read the docs.
  • [FEATURE] Statsd: Now supports multiple metrics per packet.
  • [FEATURE] Some under the hood work to support more platforms.
  • [FEATURE] Bug fixes
  • [BUGFIX] Fixes invalid configuration parsing in the case of pure JVM metrics.




  • [BUGFIX] Fixes invalid configuration parsing in the case of pure JVM metrics.




  • [FEATURE] Dogstream (parsing logs with dd-agent) supports parsing classes in addition to parsing functions.




  • [FEATURE] Multi-instance JMX check
  • [FEATURE] dogstatsd counters now send 0 for up to 10 minutes after the last increment(). They work with alerts.
  • [BUGFIX] [part 1 of #165]( dogstatsd's average is fixed
  • [BUGFIX] HAProxy logging level was logging debug messages by default.




  • [FEATURE] Deploy Pup along with the Agent (though Pup doesn't run on CentOS 5)
  • [FEATURE] Added a one line install script
  • [FEATURE] HAProxy integration
  • [BUGFIX] Run the Agent on Redhat reboots
  • [BUGFIX] #150 - Fix Pexpect dependency
  • [BUGFIX] Small fixes to the HAProxy and Elastic Search integrations.
  • [BUGFIX] Fixed a couple of host aliasing issues.


  • This version depends on Supervisor version 3 instead of version 2.3.
  • changeset




  • [BUGFIX] Incorrect version dependencies between datadog-agent and datadog-agent-base.
  • [BUGFIX] #130 Fixes a crash when changing the listening port of the forwarder.

How to upgrade from 3.0.4 on Debian and Ubuntu

When we released datadog-agent 3.0.4 we made a mistake and messed up the version dependency between packages. As a result, whenever you run apt-get upgrade or apt-get dist-upgrade and 3.0.4 is installed you may get the following error:

E: Couldn't configure pre-depend datadog-agent-base for datadog-agent, probably a dependency cycle.

If that's the case, don't panic: there is a simple fix. Simply run:

sudo apt-get update
sudo apt-get remove datadog-agent
sudo apt-get install datadog-agent

to get the new versions up-and-running.




  • [FEATURE] #112 Thanks to @charles-dyfis-net, the agent supports extra emitters. An emitter is an output for events and metrics.
  • [FEATURE] #117 Thanks to @rl-0x0, the agent can now parse supervisord logs and turn them into events and metrics.
  • [FEATURE] #121 Thanks to @charles-dyfis-net, the agent supports custom checks. Check out our README for more details.




  • [BUGFIX] #82 Now proudly runs on Amazon Web Services Linux.
  • [FEATURE] #110 More ElasticSearch metrics




  • [BUGFIX] #105 Fix for ElasticSearch 0.18



  • [BUGFIX] Support for ElasticSearch 0.18
  • [BUGFIX] #95 Fix for incorrect supervisord configuration on debian. More details here to test whether you are affected.


  • This is a major version


  • [FEATURE] dogstatsd, a drop-in replacement of statsd with tagging magic, bundled with the agent. Compatible with all statsd clients.
  • [FEATURE] #21 Support for ElasticSearch 0.19



  • [FEATURE] #83 Support authenticated connections to redis. Simply use the following stanza in /etc/dd-agent/datadog.conf to support multiple servers

    redis_urls: host:port, password@host:port




  • [BUGFIX] #80 Agent now runs happily in locales with different radix separator
  • [FEATURE] #73 Allow checks to record tag metrics





  • [BUGFIX] #68 Fixes off-by-one dog parsing error
  • [BUGFIX] #65 More robust uninstall script on Debian & Ubuntu
  • [FEATURE] #71 Supports for network metrics on Mac OS X
  • [FEATURE] #62 Sends instance-related EC2 metadata to Datadog to enable host aliases




  • [BUGFIX] fixes used memory metric
  • [BUGFIX] fixes mongo support on Ubuntu 11.10 (with pymongo 1.11)
  • [BUGFIX] fixes IO metrics on Ubuntu 12.04 (thanks @dcrosta)



  • [FEATURE] Supports Varnish 2.x


If you use ganglia, you want this version.


  • [PERFORMANCE] major ganglia speedup, getting telnetlib out of the equation




  • [BUGFIX] fixes MongoDB support, broken in 2.2.19.




  • [BUGFIX] varnish support is now xml-based, to not break when reading bitmap values (#42)
  • [BUGFIX] less verbose errors in dogstream (#55)
  • [FEATURE] Now capturing master/slave changes in Mongo



  • [BUGFIX] When using use_ec2_instance_id: yes on a non-ec2 box, don't hang! (introduced with 2.2.17)
  • [FEATURE] Initial Varnish support



  • [BUGFIX] On CentOS, pid was always saved in /tmp/ (#51)
  • CONFIGURATION CHANGE: When running on EC2, the instance id will be used in lieu of the hostname, unless use_ec2_instance_id is set no no.



  • [FEATURE] Agent auto-detects the fact that it is running on Amazon EC2
  • [FEATURE] Agent supports custom event parsers




  • [BUGFIX] Fixes MongoDB configuration parsing.




  • [BUGFIX] Used memory was not reported on 2.2.12 when running the agent on Debian Lenny.
  • [BUGFIX] Cacti memory is reported in MB, not in bytes.




  • [BUGFIX] Cacti check should fail gracefully if it cannot find RRD files.




  • [BUGFIX] Prevent counters from wrapping (#23)
  • [BUGFIX] Collect shared memory metric, accessible in Datadog via system.mem.shared.




  • [BUGFIX] On CentOS5, when both datadog-agent and datadog-agent-base are installed, datadog-agent-base runs with the stock 2.4 python, which allows python modules that support integrations (e.g. mysql) to be installed with yum.

2.2.9 (minor)



  • [FEATURE] Added support for cacti
  • [FEATURE] Added support for new memory metrics: system.mem.buffers, system.mem.cached, system.mem.buffers, system.mem.usable,