Repository navigation
Releases
Releases · gokhanertogan/RunSignal
RunSignal 0.4.0
Added
runsignal export grafana-alerts: the alert rules as a Grafana alerting provisioning file, for Grafana-managed
alerting (Grafana Cloud, or Grafana without a Prometheus ruler). The PromQL and thresholds are the same as
export prometheus, and the uids are stable.- Workload
ownerandrunbookUrl, in every SDK and in the manifest. Every New Relic alert's description names
the owner and links the runbook. Prometheus rules get anownerlabel, so Alertmanager can route to the team,
and arunbook_urlannotation.discover kubernetesreadsrunsignal.io/ownerandrunsignal.io/runbook-url.
validatechecks them, andcompareshows owner changes. runsignal validate: checks manifests offline for CI. Errors are invalid names, duplicate workloads, and
unreadable schedules or time zones. Warnings are expectations a backend cannot alert on, naming the backend.
It exits with 1 on errors, or on warnings too with--strict.runsignal compare: the alerts a manifest change adds, removes or changes, offline, in reader terms such as
"a run takes longer than 30m" → "45m".--markdownwrites it for a pull request comment.- GitHub Action:
pr-commentcomments on pull requests with the alert changes against the base branch and the
validation findings, updates the same comment on every push, and fails when a manifest has errors.commandis
now optional, andcli-pathuses an existingrunsignalinstead of installing one. - Dashboards: a panel or widget for the oldest unfinished run (
active_run_age, whathungalerts on) on the
Grafana and New Relic dashboards. The New Relic workload table shows each workload's last success as a date.
Fixed
- Grafana dashboard: per-run workloads (
runsignal exec,discover kubernetes) showed no runs, items or durations,
because each run is its own series andincrease()cannot see it grow. Their series are now summed per run, and
their last success comes from successful runs.
RunSignal 0.3.0
Added
- The CLI as a native executable of about 10 MB with no .NET dependency, for images and machines without .NET:
- release assets for Linux (glibc 2.28+ and musl, x64 and arm64), macOS and Windows;
- the container image
ghcr.io/gokhanertogan/runsignaltoCOPY --from(/runsignal,/runsignal-musl).
runsignal exec: runs any command (a shell script, a crontab entry, a container in a Kubernetes CronJob) as one
run of a workload and exports it over OTLP. Exit code 0 is a success, other codes fail witherror.type
exit_code_<n>, and stop signals are forwarded and count as cancelled. The command's exit code is returned.runsignal discover kubernetes: writes a manifest from Kubernetes CronJobs, with their schedules, time zones and
activeDeadlineSeconds, plusrunsignal.io/*annotations for severity, max duration, minimum output, failure
rate, workload name and opting out.- Manifest:
"reporting": "per-run"for workloads whose runs are short-lived processes.missedthen means no
successful run within the schedule plus grace, and there is no not-reporting or stalled alert. On New Relic,
per-runmissedneeds a run at least every 48 hours. - Manifest: a per-workload
timeZonefor its schedule. - Hung runs: a new
runsignal.workload.active_run_agegauge in every SDK reports how long the oldest unfinished
run has been running (0 when none is). It is reported for workloads that havemaxDurationor have started a
run. New Relic gets ahungcondition and Prometheus aRunSignalHungrule, which open when an unfinished run
passesmaxDuration. Before, a run that never finished only showed up later asmissedorstalled. The
Prometheus rule reads the gauge directly, so unlikeoverrunit also covers limits above the largest histogram
bucket. - .NET:
WorkloadRun.Discard()forgets a started run without recording it. The Quartz integration uses it for
vetoed jobs.
Fixed
runsignal execno longer crashes after the command has run when it exports over HTTPS in an image without
OpenSSL; it warns, skips the export and returns the command's exit code.- Alert descriptions formatted durations with the machine's culture ("1,5h" on a Turkish machine, "1.5h" in CI),
soprovisionfrom two machines kept updating the same conditions. They are now culture-invariant. - Python: a second
WorkloadRegistryon the same meter provider (for example one you create next to
default_registry) silently lost its gauges, so its workloads never gotmissed,stalledorhungalerts.
Registries on one provider now share their gauges. - Node.js: a workload in
defaultRegistrythat never runs reports its gauges once the SDK has started, so it gets
itsmissedalert. Before, the gauges moved to the SDK's meter provider only on the first recording. - Java: a registered workload reports its gauges at startup, so a cron job that never fires gets its
missed
alert.WorkloadRegistry.global()does this only onceGlobalOpenTelemetryis set (as the Java agent does), so
it never fixes the global instance to a no-op. - Java (Spring Batch 6): when the final job repository update fails, Spring Batch skips
afterJob, which left the
run in progress. A background sweep now ends such runs as failed witherror.type
job_repository_update_failed, about 30 to 60 seconds after the job ends. - Java: a reactive
@Scheduledmethod whose subscription is cancelled is recorded as cancelled, not successful. - Go: a run whose function calls
runtime.Goexit(ast.FailNowdoes) fails witherror.typegoexitinstead
of staying in progress. - Rust: a registered workload reports its gauges straight away, so a scheduled workload that never runs gets its
missedalert. Before, gauges appeared only after the first recording.Registry::global()follows the global
meter provider until it first collects. If workloads are registered before the provider is set and record
nothing, call the newRegistry::refresh_meter_provider()once after setting it. - Python (Celery): a task id started twice (a redelivery, or a nested eager apply) no longer loses its first run.
instrument()on a second registry now connects its signal handlers and records on that registry, instead of
silently doing nothing or recording on the first one.
Changed
- A run that is started and never ended now stays in progress, so it eventually opens
hung. Before, it recorded
nothing.
RunSignal 0.2.0
Added
- Python: the
runsignalpackage on PyPI, with Celery (tasks and beat schedules) and APScheduler 3
integrations,with/async withruns and a@monitordecorator. It passes the shared conformance vectors
and writes the same manifest as the other SDKs. - Node.js:
cronJob(@runsignal/node/cron) monitors jobs scheduled with node-cron, croner, cron or
node-schedule, and registers them with their schedule at startup so a job that never fires gets amissed
alert. - CLI:
runsignal export grafanawrites a Grafana dashboard for RunSignal metrics in Prometheus. - NuGet packages have an icon.
- A GitHub Action (
uses: gokhanertogan/RunSignal@<ref>) that installs the CLI and runs any of its commands in CI.
RunSignal 0.1.0
First release.
Signals and specification
- The same OpenTelemetry metrics in every SDK: runs, duration, items, heartbeats, and the
since_successand
since_heartbeatgauges. Runs are also recorded as spans. - A workload manifest format (
spec/manifest.schema.json) and cross-language conformance vectors for naming
and signals.
SDKs and integrations
- .NET:
RunSignal.Core,RunSignal.DotNet,RunSignal.OpenTelemetry, plus Hangfire 1.8+ (RunSignal.Hangfire)
and Quartz.NET 3 and 4 (RunSignal.Quartz,RunSignal.Quartz4). - Go:
github.com/gokhanertogan/RunSignal/go, with robfig/cron (rscron). - Java:
io.github.gokhanertogan.runsignal:runsignalandrunsignal-spring-boot-starterfor Spring
@Scheduledand Spring Batch 5 and 6. - Node.js:
@runsignal/node, with BullMQ. - Rust:
runsignal.
Alerting
runsignalCLI:plan,diff,provisionanddestroyfor New Relic alert policies, NRQL conditions,
workflows and dashboards, andexport prometheusfor Prometheus alerting rules.- Alerts: missed, failed, stalled, overrun, low output, failure rate, item failure rate and not reporting.