Perfkit Benchmarker contains set of benchmarks to measure and compare cloud offerings. The benchmarks use defaults to reflect what most users will see. PerfKit Benchmarker is licensed under the Apache 2 license terms. Please make sure to read, understand and agree to the terms of the LICENSE and CONTRIBUTING files before proceeding.
tohaowu and gareth-ferneyhough Clean redundant code in TPU test.
Created by MOE:
Latest commit 333b44e Sep 25, 2018

PerfKit Benchmarker

PerfKit Benchmarker is an open effort to define a canonical set of benchmarks to measure and compare cloud offerings. It's designed to operate via vendor provided command line tools. The benchmark default settings are not tuned for any particular platform or instance type. These settings are recommended for consistency across services. Only in the rare case where there is a common practice like setting the buffer pool size of a database do we change any settings.

This README is designed to give you the information you need to get running with the benchmarker and the basics of working with the code. The wiki contains more detailed information:

Known Issues


PerfKit Benchmarker provides wrappers and workload definitions around popular benchmark tools. We made it very simple to use and automate everything we can. It instantiates VMs on the Cloud provider of your choice, automatically installs benchmarks, and runs the workloads without user interaction.

Due to the level of automation you will not see prompts for software installed as part of a benchmark run. Therefore you must accept the license of each of the benchmarks individually, and take responsibility for using them before you use the PerfKit Benchmarker.

In its current release these are the benchmarks that are executed:

Some of the benchmarks invoked require Java. You must also agree with the following license:

CoreMark setup cannot be automated. EEMBC requires users to agree with their terms and conditions, and PerfKit Benchmarker users must manually download the CoreMark tarball from their website and save it under the perfkitbenchmarker/data folder (e.g. ~/PerfKitBenchmarker/perfkitbenchmarker/data/coremark_v1.0.tgz)

SPEC CPU2006 benchmark setup cannot be automated. SPEC requires that users purchase a license and agree with their terms and conditions. PerfKit Benchmarker users must manually download cpu2006-1.2.iso from the SPEC website, save it under the perfkitbenchmarker/data folder (e.g. ~/PerfKitBenchmarker/perfkitbenchmarker/data/cpu2006-1.2.iso), and also supply a runspec cfg file (e.g. ~/PerfKitBenchmarker/perfkitbenchmarker/data/linux64-x64-gcc47.cfg). Alternately, PerfKit Benchmarker can accept a tar file that can be generated with the following steps:

  • Extract the contents of cpu2006-1.2.iso into a directory named cpu2006
  • Run cpu2006/
  • Copy the cfg file into cpu2006/config
  • Create a tar file containing the cpu2006 directory, and place it under the perfkitbenchmarker/data folder (e.g. ~/PerfKitBenchmarker/perfkitbenchmarker/data/cpu2006v1.2.tgz).

PerfKit Benchmarker will use the tar file if it is present. Otherwise, it will search for the iso and cfg files.

Installing PerfKit Benchmarker and Prerequisites

Before you can run the PerfKit Benchmarker, you need account(s) on the cloud provider(s) you want to benchmark:

You also need the software dependencies, which are mostly command line tools and credentials to access your accounts without a password. The following steps should help you get the CLI tool auth in place.

If you are running on Windows, you will need to install GitHub Windows since it includes tools like openssl and an ssh client. Alternatively, you can install Cygwin since it should include the same tools.

Install Python 2.7 and pip

If you are running on Windows, get the latest version of Python 2.7 here. This should have pip bundled with it. Make sure your PATH environment variable is set so that you can use both python and pip on the command line (you can have the installer do it for you if you select the correct option).

Most Linux distributions and recent Mac OS X versions already have Python 2.7 installed. If Python is not installed, you can likely install it using your distribution's package manager, or see the Python Download page.

If you need to install pip, see these instructions.

(Windows Only) Install GitHub Windows


Make sure that openssl/ssh/scp/ssh-keygen are on your path (you will need to update the PATH environment variable). The path to these commands should be


Install PerfKit

Download PerfKit Benchmarker from GitHub.

Install PerfKit Benchmarker dependencies

$ cd /path/to/PerfKitBenchmarker
$ sudo pip install -r requirements.txt

Preprovisioned data

Some benchmarks may require data to be preprovisioned in a cloud. To preprovision data, you will need to obtain the data and then upload it to that cloud. See more information below about which benchmarks require preprovisioned data and how to upload it to different clouds.

Note. Before we start to switch over to preprovisioned data, we should support a fallback strategy of downloading files to the data/ directory on the machine used to run PerfKitBenchmarker (as is done today for CoreMark and SPEC CPU2006).

Cloud account setup

This section describes the setup steps needed for each cloud system. Note that you only need to perform setup steps on the clouds you wish to test. If you only want to test Google Cloud, you only need to install and configure gcloud.

After configuring the clouds you intend to use, skip to Running a Single Benchmark, unless you are going to use an object storage benchmark, in which case you need to configure a boto file.

Install gcloud and setup authentication

Instructions: If you're using OS X or Linux, you can run the command below:

$ curl | bash

When prompted, pick the local folder, then Python project, then the defaults for all the rest.

Restart your shell window (or logout/ssh again if running on a VM)

On Windows, visit the same page and follow the Windows installation instructions on the page.

Next, create a project by visiting Google Cloud Console. After that, run:

$ gcloud init

which helps you authenticate, set your project, and set some defaults.

Alternatively, if that is already set up, and you just need to authenticate, you can use:

$ gcloud auth login

For help, see gcloud docs.

Install OpenStack CLI client and setup authentication

Make sure you have installed pip (see the section above).

Install OpenStack CLI utilities via the following command:

$ sudo pip install -r perfkitbenchmarker/providers/openstack/requirements.txt

To setup credentials and endpoint information simply set the environment variables using an OpenStack RC file. For help, see OpenStack docs

Kubernetes configuration and credentials

Perfkit uses the kubectl binary in order to communicate with a Kubernetes cluster - you need to pass the path to the kubectl binary using the --kubectl flag. It's recommended to use version 1.0.1. Authentication to a Kubernetes cluster is done via a kubeconfig file. Its path is passed using the --kubeconfig flag.

Image prerequisites
Please refer to the Image prerequisites for Docker based clouds.

Kubernetes cluster configuration
If your Kubernetes cluster is running on CoreOS:

  1. Fix $PATH environment variable so that the appropriate binaries can be found:

    $ sudo mkdir /etc/systemd/system/kubelet.service.d
    $ sudo vim /etc/systemd/system/kubelet.service.d/10-env.conf

    Add the following line to the [Service] section:

  2. Reboot the node:

    $ sudo reboot

Note that some benchmarks must be run within a privileged container. By default Kubernetes doesn't allow containers to be scheduled in privileged mode - you have to add the --allow-privileged=true flag to kube-apiserver and each kubelet startup command.

Ceph integration
When you run benchmarks with the standard scratch disk type (--scratch_disk_type=standard - which is a default option), Ceph storage will be used. There are some configuration steps you need to follow before you will be able to spawn Kubernetes PODs with Ceph volume. On each Kubernetes node, and on the machine which is running the Perfkit benchmarks, do the following:

  1. Copy /etc/ceph directory from Ceph-host.

  2. Install ceph-common package so that rbd command is available:

  • If your Kubernetes cluster is running on CoreOS, then you need to create a bash script called rbd which will run the rbd command inside a Docker container:

    /usr/bin/docker run -v /etc/ceph:/etc/ceph -v /dev:/dev -v /sys:/sys  --net=host --privileged=true --rm=true ceph/rbd $@

    Save the file as rbd and run:

    $ chmod +x rbd
    $ sudo mkdir /opt/bin
    $ sudo cp rbd /opt/bin

    Install rbdmap:

    $ git clone
    $ cd ceph-docker/examples/coreos/rbdmap/
    $ sudo mkdir /opt/sbin
    $ sudo cp rbdmap /opt/sbin
    $ sudo cp ceph-rbdnamer /opt/bin
    $ sudo cp 50-rbd.rules /etc/udev/rules.d
    $ sudo reboot

You have two Ceph authentication options available:

  1. Keyring - pass the path to the keyring file using --ceph_keyring flag

  2. Secret. In this case you have to create a secret first:

    Retrieve base64-encoded Ceph admin key:

    $ ceph auth get-key client.admin | base64

    Create a file called create_ceph_admin.yml and replace the key value with the output from the previous command:

    apiVersion: v1
    kind: Secret
      name: my-ceph-secret

    Add secret to Kubernetes:

    $ kubectl create -f create_ceph_admin.yml

    You will have to pass the Secret name (using --ceph_secret flag) when running the benchmakrs. In this case it should be: --ceph_secret=my-ceph-secret.

Mesos configuration

Mesos provider communicates with Marathon framework in order to manage Docker instances. Thus it is required to setup Marathon framework along with the Mesos cluster. In order to connect to Mesos you need to provide IP address and port to Marathon framework using --marathon_address flag.

Provider has been tested with Mesos v0.24.1 and Marathon v0.11.1.

Overlay network
Mesos on its own doesn't provide any solution for overlay networking. You need to configure your cluster so that the instances will live in the same network. For this purpose you may use Flannel, Calico, Weave, etc.

Mesos cluster configuration
Make sure your Mesos-slave nodes are reachable (by hostname) from the machine which is used to run the benchmarks. In case they are not, edit the /etc/hosts file appropriately.

Image prerequisites
Please refer to the Image prerequisites for Docker based clouds.

Cloudstack: Install dependencies and set the API keys

$ sudo pip install -r perfkitbenchmarker/providers/cloudstack/requirements.txt

Get the API key and SECRET from Cloudstack. Set the following environement variables.

export CS_API_URL=<insert API endpoint>
export CS_API_KEY=<insert API key>
export CS_API_SECRET=<insert API secret>

Specify the network offering when running the benchmark. If using VPC (--cs_use_vpc), also specify the VPC offering (--cs_vpc_offering).

$ ./ --cloud=CloudStack --benchmarks=ping --cs_network_offering=DefaultNetworkOffering

Install AWS CLI and setup authentication

Make sure you have installed pip (see the section above).

Follow instructions at or run the following command (omit the 'sudo' on Windows)

$ sudo pip install -r perfkitbenchmarker/providers/aws/requirements.txt

Navigate to the AWS console to create access credentials:

  • On the console click on your name (top left)
  • Click on "Security Credentials"
  • Click on "Access Keys", the create New Access Key. Download the file, it contains the Access key and Secret keys to access services. Note the values and delete the file.

Configure the CLI using the keys from the previous step:

$ aws configure

Windows Azure CLI and credentials

You first need to install node.js and NPM. This version of Perfkit Benchmarker is known to be compatible with Azure CLI version 0.10.4, and will likely work with any version newer than that.

Go here, and follow the setup instructions.

Next, run the following (omit the sudo on Windows):

$ sudo npm install azure-cli -g
$ azure login

Test that azure is installed correctly:

$ azure vm list

Finally, make sure Azure is in Resource Management mode and that your account is authorized to allocate VMs and networks from Azure:

$ azure config mode arm
$ azure provider register Microsoft.Compute
$ azure provider register Microsoft.Network

Install AliCloud CLI and setup authentication

Make sure you have installed pip (see the section above).

Run the following command to install aliyuncli (omit the sudo on Windows)

  1. Install python development tools:

    In Debian or Ubuntu:

    $ sudo apt-get install -y python-dev

    In CentOS:

    $ sudo yum install python-devel
  2. Install aliyuncli tool and python SDK for ECS:

    $ sudo pip install -r perfkitbenchmarker/providers/alicloud/requirements.txt

    In some CentOS version, you may need:

    $ sudo yum install libffi-devel.x86_64
    $ sudo yum install openssl-devel.x86_64
    $ sudo pip install 'colorama<=0.3.3'

    To check if AliCloud is installed:

    $ aliyuncli --help

    Check if aliyuncli ecs command is ready:

    $ aliyuncli ecs help

    If you see the "usage" message, you should follow step 3. Otherwise, jump to step 4.

  3. Dealing with an exception when it runs on some specific version of Ubuntu. Get the python lib path: /usr/lib/python2.7/dist-packages

    $ python
    > from distutils.sysconfig import get_python_lib
    > get_python_lib()

    Copy to the right directory (for Python 2.7.X):

    $ sudo cp -r /usr/local/lib/python2.7/dist-packages/aliyun* /usr/lib/python2.7/dist-packages/

    Check again:

    $ aliyuncli ecs help
  4. Navigate to the AliCloud console to create access credentials:

    • Login first
    • Click on "AccessKeys" (top right)
    • Click on "Create Access Key", copy and store the "Access Key ID" and "Access Key Secret" to a safe place.
    • Configure the CLI using the Access Key ID and Access Key Secret from the previous step
    $ aliyuncli configure

DigitalOcean configuration and credentials

  1. Install doctl, the DigitalOcean CLI, following the instructions at

  2. Authenticate with doctl. The easiest way is running doctl auth login and following the instructions, but any of the options at the doctl site will work.

Installing CLIs and credentials for Rackspace

In order to interact with the Rackspace Public Cloud, PerfKitBenchmarker makes use of RackCLI. You can find the instructions to install and configure RackCLI here:

To run PerfKit Benchmarker against Rackspace is very easy. Simply make sure Rack CLI is installed and available in your PATH, optionally use the flag --rack_path to indicate the path to the binary.

For a Rackspace UK Public Cloud account, unless it's your default RackCLI profile then it's recommended that you create a profile for your UK account. Once configured, use flag --profile to specify which RackCLI profile to use. You can find more details here:

Note: Not all flavors are supported on every region. Always check first if the flavor is supported in the region.

ProfitBricks configuration and credentials

Get started by running:

$ sudo pip install -r perfkitbenchmarker/providers/profitbricks/requirements.txt

PerfKit Benchmarker uses the Requests module to interact with ProfitBricks' REST API. HTTP Basic authentication is used to authorize access to the API. Please set this up as follows:

Create a configuration file containing the email address and password associated with your ProfitBricks account, separated by a colon. Example:

$ less ~/.config/profitbricks-auth.cfg

The PerfKit Benchmarker will automatically base64 encode your credentials before making any calls to the REST API.

PerfKit Benchmarker uses the file location ~/.config/profitbricks-auth.cfg by default. You can use the --profitbricks_config flag to override the path.

Image prerequisites for Docker based clouds

Docker instances by default don't allow to SSH into them. Thus it is important to configure your Docker image so that it has SSH server installed. You can use your own image or build a new one based on a Dockerfile placed in tools/docker_images directory - in this case please refer to Docker images document.

Create and configure a .boto file for object storage benchmarks

In order to run object storage benchmark tests, you need to have a properly configured ~/.boto file. The directions require that you have installed google-cloud-sdk. The directions for doing that are in the gcloud installation section.

Here is how:

  • Create the ~/.boto file (If you already have ~/.boto, you can skip this step. Consider making a backup copy of your existing .boto file.)

To create a new ~/.boto file, issue the following command and follow the instructions given by this command:

$ gsutil config

As a result, a .boto file is created under your home directory.

Open the .boto file and edit the following fields:

  1. In the [Credentials] section:

    gs_oauth2_refresh_token: set it to be the same as the refresh_token field in your gcloud credential file (~/.config/gcloud/credentials.db), which was setup as part of the gcloud auth login step. To see the refresh token, run

    $ strings ~/.config/gcloud/credentials.db.

    aws_access_key_id, aws_secret_access_key: set these to be the AWS access keys you intend to use for these tests, or you can use the same keys as those in your existing AWS credentials file (~/.aws/credentials).

  2. In the [GSUtil] section:

    default_project_id: if it is not already set, set it to be the google cloud storage project ID you intend to use for this test. (If you used gsutil config to generate the .boto file, you should have been prompted to supply this information at this step).

  3. In the [OAuth2] section:

    client_id, client_secret: set these to be the same as those in your gcloud credentials file (~/.config/gcloud/credentials.db), which was setup as part of the gcloud auth login step.

Running a Single Benchmark

PerfKit Benchmarker can run benchmarks both on Cloud Providers (GCP, AWS, Azure, DigitalOcean) as well as any "machine" you can SSH into.

Example run on GCP

$ ./ --project=<GCP project ID> --benchmarks=iperf --machine_type=f1-micro

Example run on AWS

$ cd PerfKitBenchmarker
$ ./ --cloud=AWS --benchmarks=iperf --machine_type=t2.micro

Example run on Azure

$ ./ --cloud=Azure --machine_type=Standard_A0 --benchmarks=iperf

Example run on AliCloud

$ ./ --cloud=AliCloud --machine_type=ecs.s2.large --benchmarks=iperf

Example run on DigitalOcean

$ ./ --cloud=DigitalOcean --machine_type=16gb --benchmarks=iperf

Example run on OpenStack

$ ./ --cloud=OpenStack --machine_type=m1.medium \
           --openstack_network=private --benchmarks=iperf

Example run on Kubernetes

$ ./ --cloud=Kubernetes --benchmarks=iperf --kubectl=/path/to/kubectl --kubeconfig=/path/to/kubeconfig --image=image-with-ssh-server  --ceph_monitors=,

Example run on Mesos

$ ./ --cloud=Mesos --benchmarks=iperf --marathon_address=localhost:8080 --image=image-with-ssh-server

Example run on CloudStack

./ --cloud=CloudStack --benchmarks=ping --cs_network_offering=DefaultNetworkOffering

Example run on Rackspace

$ ./ --cloud=Rackspace --machine_type=general1-2 --benchmarks=iperf

Example run on ProfitBricks

$ ./ --cloud=ProfitBricks --machine_type=Small --benchmarks=iperf

How to Run Windows Benchmarks

Install all dependencies as above and ensure that smbclient is installed on your system if you are running on a linux controller:

$ which smbclient

Now you can run Windows benchmarks by running with --os_type=windows. Windows has a different set of benchmarks than Linux does. They can be found under perfkitbenchmarker/windows_benchmarks/. The target VM OS is Windows Server 2012 R2.

How to Run Benchmarks with Juju

Juju is a service orchestration tool that enables you to quickly model, configure, deploy and manage entire cloud environments. Supported benchmarks will deploy a Juju-modeled service automatically, with no extra user configuration required, by specifying the --os_type=juju flag.


$ ./ --cloud=AWS --os_type=juju --benchmarks=cassandra_stress

Benchmark support

Benchmark/Package authors need to implement the JujuInstall() method inside their package. This method deploys, configures, and relates the services to be benchmarked. Please note that other software installation and configuration should be bypassed when FLAGS.os_type == JUJU. See perfkitbenchmarker/linux_packages/ for an example implementation.

How to Run All Standard Benchmarks

Run without the --benchmarks parameter and every benchmark in the standard set will run serially which can take a couple of hours (alternatively, run with --benchmarks="standard_set"). Additionally, if you don't specify --cloud=..., all benchmarks will run on the Google Cloud Platform.

How to Run All Benchmarks in a Named Set

Named sets are are groupings of one or more benchmarks in the benchmarking directory. This feature allows parallel innovation of what is important to measure in the Cloud, and is defined by the set owner. For example the GoogleSet is maintained by Google, whereas the StanfordSet is managed by Stanford. Once a quarter a meeting is held to review all the sets to determine what benchmarks should be promoted to the standard_set. The Standard Set is also reviewed to see if anything should be removed. To run all benchmarks in a named set, specify the set name in the benchmarks parameter (e.g., --benchmarks="standard_set"). Sets can be combined with individual benchmarks or other named sets.

Useful Global Flags

The following are some common flags used when configuring PerfKit Benchmarker.

Flag Notes
--helpmatch=pkb see all global flags
--helpmatch=hpcc see all flags associated with the hpcc benchmark. You can substitute any benchmark name to see the associated flags.
--benchmarks A comma separated list of benchmarks or benchmark sets to run such as --benchmarks=iperf,ping . To see the full list, run ./ --helpmatch=benchmarks | grep perfkitbenchmarker
--cloud Cloud where the benchmarks are run. See the table below for choices.
--machine_type Type of machine to provision if pre-provisioned machines are not used. Most cloud providers accept the names of pre-defined provider-specific machine types (for example, GCP supports --machine_type=n1-standard-8 for a GCE n1-standard-8 VM). Some cloud providers support YAML expressions that match the corresponding VM spec machine_type property in the YAML configs (for example, GCP supports --machine_type="{cpus: 1, memory: 4.5GiB}" for a GCE custom VM with 1 vCPU and 4.5GiB memory). Note that the value provided by this flag will affect all provisioned machines; users who wish to provision different machine types for different roles within a single benchmark run should use the YAML configs for finer control.
--zones This flag allows you to override the default zone. See the table below.
--data_disk_type Type of disk to use. Names are provider-specific, but see table below.

The default cloud is 'GCP', override with the --cloud flag. Each cloud has a default zone which you can override with the --zones flag, the flag supports the same values that the corresponding Cloud CLIs take:

Cloud name Default zone Notes
GCP us-central1-a
AWS us-east-1a
Azure East US
AliCloud West US
DigitalOcean sfo1 You must use a zone that supports the features 'metadata' (for cloud config) and 'private_networking'.
OpenStack nova
CloudStack QC-1
Rackspace IAD OnMetal machine-types are available only in IAD zone
Kubernetes k8s
ProfitBricks AUTO Additional zones: ZONE_1, ZONE_2, or ZONE_3


./ --cloud=GCP --zones=us-central1-a --benchmarks=iperf,ping

The disk type names vary by provider, but the following table summarizes some useful ones. (Many cloud providers have more disk types beyond these options.)

Cloud name Network-attached SSD Network-attached HDD
GCP pd-ssd pd-standard
AWS gp2 standard
Azure Premium_LRS Standard_LRS
Rackspace cbs-ssd cbs-sata

Also note that --data_disk_type=local tells PKB not to allocate a separate disk, but to use whatever comes with the VM. This is useful with AWS instance types that come with local SSDs, or with the GCP --gce_num_local_ssds flag.

If an instance type comes with more than one disk, PKB uses whichever does not hold the root partition. Specifically, on Azure, PKB always uses /dev/sdb as its scratch device.

Proxy configuration for VM guests.

If the VM guests do not have direct Internet access in the cloud environment, you can configure proxy settings through flags.

To do that simple setup three flags (All urls are in notation ): The flag values use the same <protocol>://<server>:<port> syntax as the corresponding environment variables, for example --http_proxy= .

Flag Notes
--http_proxy Needed for package manager on Guest OS and for some Perfkit packages
--https_proxy Needed for package manager or Ubuntu guest and for from github downloaded packages
--ftp_proxy Needed for some Perfkit packages

Preprovisioned Data

As mentioned above, some benchmarks require preprovisioned data. This section describes how to preprovision this data.

Benchmarks with Preprovisioned Data

Sample Preprovision Benchmark

This benchmark demonstrates the use of preprovisioned data. Create the following file to upload using the command line:

echo "1234567890" > preprovisioned_data.txt

To upload, follow the instructions below with a filename of preprovisioned_data.txt and a benchmark name of sample.

Clouds with Preprovisioned Data

Google Cloud

To preprovision data on Google Cloud, you will need to upload each file to Google Cloud Storage using gsutil. First, you will need to create a storage bucket that is accessible from VMs created in Google Cloud by PKB. Then copy each file to this bucket using the command

gsutil cp <filename> gs://<bucket>/<benchmark-name>/<filename>

To run a benchmark on Google Cloud that uses the preprovisioned data, use the flag --gcp_preprovisioned_data_bucket=<bucket>.


To preprovision data on AWS, you will need to upload each file to S3 using the AWS CLI. First, you will need to create a storage bucket that is accessible from VMs created in AWS by PKB. Then copy each file to this bucket using the command

aws s3 cp <filename> s3://<bucket>/<benchmark-name>/<filename>

To run a benchmark on AWS that uses the preprovisioned data, use the flag --aws_preprovisioned_data_bucket=<bucket>.

Configurations and Configuration Overrides

Each benchmark now has an independent configuration which is written in YAML. Users may override this default configuration by providing a configuration. This allows for much more complex setups than previously possible, including running benchmarks across clouds.

A benchmark configuration has a somewhat simple structure. It is essentially just a series of nested dictionaries. At the top level, it contains VM groups. VM groups are logical groups of homogenous machines. The VM groups hold both a vm_spec and a disk_spec which contain the parameters needed to create members of that group. Here is an example of an expanded configuration:

      vm_count: 4
          machine_type: n1-standard-1
          image: ubuntu-14-04
          zone: us-central1-c
          machine_type: m3.medium
          image: ami-######
          zone: us-east-1a
        # Other clouds here...
      # This specifies the cloud to use for the group. This allows for
      # benchmark configurations that span clouds.
      cloud: AWS
      # No disk_spec here since these are loaders.
      vm_count: 1
      cloud: GCP
            cpus: 2
            memory: 10.0GiB
          image: ubuntu-14-04
          zone: us-central1-c
        # Other clouds here...
      disk_count: 1
          disk_size: 100
          disk_type: standard
          mount_point: /scratch
        # Other clouds here...
      vm_count: 4
      cloud: GCP
          machine_type: n1-standard-4
          image: ubuntu-14-04
          zone: us-central1-c
        # Other clouds here...
      disk_count: 1
          disk_size: 500
          disk_type: remote_ssd
          mount_point: /scratch
        # Other clouds here...

For a complete list of keys for vm_specs and disk_specs see virtual_machine.BaseVmSpec and disk.BaseDiskSpec and their derived classes.

User configs are applied on top of the existing default config and can be specified in two ways. The first is by supplying a config file via the --benchmark_config_file flag. The second is by specifying a single setting to override via the --config_override flag.

A user config file only needs to specify the settings which it is intended to override. For example if the only thing you want to do is change the number of VMs for the cluster_boot benchmark, this config is sufficient:

      vm_count: 100

You can achieve the same effect by specifying the --config_override flag. The value of the flag should be a path within the YAML (with keys delimited by periods), an equals sign, and finally the new value:


See the section below for how to use static (i.e. pre-provisioned) machines in your config.

Advanced: How To Run Benchmarks Without Cloud Provisioning (e.g., local workstation)

It is possible to run PerfKit Benchmarker without running the Cloud provisioning steps. This is useful if you want to run on a local machine, or have a benchmark like iperf run from an external point to a Cloud VM.

In order to do this you need to make sure:

  • The static (i.e. not provisioned by PerfKit Benchmarker) machine is ssh'able
  • The user PerfKitBenchmarker will login as has 'sudo' access. (*** Note we hope to remove this restriction soon ***)

Next, you will want to create a YAML user config file describing how to connect to the machine as follows:

  - &vm1 # Using the & character creates an anchor that we can
         # reference later by using the same name and a * character.
    user_name: voellm
    ssh_private_key: /home/voellm/perfkitkeys/my_key_file.pem
    zone: Siberia
      - mount_point: /data_dir
  • The ip_address is the address where you want benchmarks to run.
  • ssh_private_key is where to find the private ssh key.
  • zone can be anything you want. It is used when publishing results.
  • disk_specs is used by all benchmarks which use disk (i.e., fio, bonnie++, many others).

In the same file, configure any number of benchmarks (in this case just iperf), and reference the static VM as follows:

        - *vm1

I called my file iperf.yaml and used it to run iperf from Siberia to a GCP VM in us-central1-f as follows:

$ ./ --benchmarks=iperf --machine_type=f1-micro --benchmark_config_file=iperf.yaml --zones=us-central1-f --ip_addresses=EXTERNAL
  • ip_addresses=EXTERNAL tells PerfKit Benchmarker not to use 10.X (ie Internal) machine addresses that all Cloud VMs have. Just use the external IP address.

If a benchmark requires two machines like iperf, you can have two machines in the same YAML file as shown below. This means you can indeed run between two machines and never provision any VMs in the Cloud.

  - &vm1
    ip_address: <ip1>
    user_name: connormccoy
    ssh_private_key: /home/connormccoy/.ssh/google_compute_engine
    install_packages: false
  - &vm2
    ip_address: <ip2>
    user_name: connormccoy
    ssh_private_key: /home/connormccoy/.ssh/google_compute_engine
    ssh_port: 2222

        - *vm2
        - *vm1

Specifying Flags in Configuration Files

You can now specify flags in configuration files by using the flags key at the top level in a benchmark config. The expected value is a dictionary mapping flag names to their new default values. The flags are only defaults; it's still possible to override them via the command line. It's even possible to specify conflicting values of the same flag in different benchmarks:

    machine_type: n1-standard-2
    zone: us-central1-b
    iperf_sending_thread_count: 2

    machine_type: n1-standard-8

The new defaults will only apply to the benchmark in which they are specified.

Using Elasticsearch Publisher

PerfKit data can optionally be published to an Elasticsearch server. To enable this, the elasticsearch Python package must be installed.

$ sudo pip install elasticsearch

Note: The elasticsearch Python library and Elasticsearch must have matching major versions.

The following are flags used by the Elasticsearch publisher. At minimum, all that is needed is the --es_uri flag.

Flag Notes
--es_uri The Elasticsearch server address and port (e.g. localhost:9200)
--es_index The Elasticsearch index name to store documents (default: perfkit)
--es_type The Elasticsearch document type (default: result)

Note: Amazon ElasticSearch service currently does not support transport on port 9200 therefore you must use endpoint with port 80 eg. search-<ID> and allow your IP address in the cluster.

Using InfluxDB Publisher

No additional packages need to be installed in order to publish Perfkit data to an InfluxDB server.

InfluxDB Publisher takes in the flags for the Influx uri and the Influx DB name. The publisher will default to the pre-set defaults, identified below, if no uri or DB name is set. However, the user is required to at the very least call the --influx_uri flag to publish data to Influx.

Flag Notes Default
--influx_uri The Influx DB address and port. Expects the format hostname:port localhost:8086
--influx_db_name The name of Influx DB database that you wish to publish to or create perfkit

How to Extend PerfKit Benchmarker

First start with the file. It has the basics on how to work with PerfKitBenchmarker, and how to submit your pull requests.

In addition to the file we have added a lot of comments into the code to make it easy to:

  • Add new benchmarks (e.g.: --benchmarks=<new benchmark>)
  • Add new package/os type support (e.g.: --os_type=<new os type>)
  • Add new providers (e.g.: --cloud=<new provider>)
  • etc.

Even with lots of comments we make to support more detailed documention. You will find the documentation we have on the wiki. Missing documentation you want? Start a page and/or open an issue to get it added.

Integration Testing

If you wish to run unit or integration tests, ensure that you have tox >= 2.0.0 installed.

In addition to regular unit tests, which are run via hooks/check-everything, PerfKit Benchmarker has integration tests, which create actual cloud resources and take time and money to run. For this reason, they will only run when the variable PERFKIT_INTEGRATION is defined in the environment. The command

$ tox -e integration

will run the integration tests. The integration tests depend on having installed and configured all of the relevant cloud provider SDKs, and will fail if you have not done so.

Planned Improvements

Many... please add new requests via GitHub issues.