Skip to content

Latest commit

 

History

History
285 lines (205 loc) · 11.2 KB

README.md

File metadata and controls

285 lines (205 loc) · 11.2 KB

Zeebe Benchmark Helm Chart

Requirements

  • Helm >= 3.11.x
  • Kubernetes >= 1.20+
  • Minimum cluster requirements include the following to run this chart with default settings. All of these settings are configurable.
    • Three Kubernetes nodes to respect the default "hard" affinity settings
    • 2GB of RAM for the JVM heap

Dependencies

Zeebe Benchmark Helm chart depends on the Camunda Platform 8 Helm chart, which has its own dependencies.

The dependency management is fully automated and managed by Helm itself; however, it's good to understand the dependency structure. This third-party dependency is reflected in the Helm chart as follows:

zeebe-benchmark
  |_camunda-platform
      |_ elasticsearch
      |_ identity
        |_ keycloak
          |_ postgresql
      |_ optimize
      |_ operate
      |_ tasklist
      |_ web-modeler
        |_ postgresql
      |_ zeebe

In the Zeebe Benchmark Helm chart, every sub-component from Camunda Platform 8 is disabled except Zeebe. This can be changed via setting the right values.

Installation

The first command adds the Zeebe Benchmark Helm charts repo, and the second installs the chart to your current Kubernetes context.

  helm repo add zeebe-benchmark https://zeebe-io.github.io/benchmark-helm/
  helm install this-is-a-benchmark zeebe-benchmark/zeebe-benchmark

Local Kubernetes

We recommend using Helm on KIND for local environments, as the Helm configurations are battle-tested and much closer to production systems.

For more details, follow the Camunda Platform 8 local Kubernetes cluster guide.

Uninstalling Charts

You can remove these charts by running:

helm uninstall YOUR_RELEASE_NAME

Note

Notice that all the Services and Pods will be deleted, but not the PersistentVolumeClaims (PVC) which are used to hold the storage for the data generated by the cluster and Elasticsearch.

To free up the storage, you need to delete all the PVCs manually.

First, view the PVCs:

kubectl get pvc -l app.kubernetes.io/instance=YOUR_RELEASE_NAME
kubectl get pvc -l release=YOUR_RELEASE_NAME

Then delete the ones that you don't want to keep:

kubectl delete pvc -l app.kubernetes.io/instance=YOUR_RELEASE_NAME
kubectl delete pvc -l release=YOUR_RELEASE_NAME

Or you can delete the related Kubernetes namespace, which contains all PVCs.

Configuration

The following sections contain the configuration values for the chart and each sub-chart. All of them can be overwritten via a separate values.yaml file.

Check out the default values.yaml file, which contains the same content and documentation.

Note For more details about deploying Camunda Platform 8 on Kubernetes, please visit the Helm/Kubernetes installation instructions docs.

Global

Section Parameter Description Default
global Global variables which can be accessed by all sub charts
image.repository Defines the repository from which to fetch the images. "gcr.io/zeebe-io"
image.tag Defines the tag / version which should be used in the chart. "SNAPSHOT"
image.pullPolicy Defines the image pull policy which should be used. Always

Camunda Platform

Zeebe Benchmark Helm chart has a dependency on the Camunda 8 Platform Helm chart. All variables related to Camunda Platform can be found in camunda-platform/values.yaml and can be set under camunda-platform.

Section Parameter Description Default
camunda-platform enabled If true, enables Camunda Platform 8 deployment as part of the Helm chart true

Example:

camunda-platform:
  enabled: true

Per default Zeebe and Zeebe-Gateway are enabled and all other components are disabled.

Retention Policy

Allows to configure the elasticsearch index retention policies, which are much more fine granular in contrast to the ones provided by the parent helm chart (camunda-platform).

Section Parameter Description Default
retentionPolicy Configuration to configure the elasticsearch index retention policies
enabled If true, elasticsearch curator cronjob and configuration will be deployed. false
schedule Defines how often/when the curator should run. "0 * * * *"
policies Defines a list of policies, which allows to specify policies for specific indexes or globally (based on the pattern specified). "0 * * * *"
image Configuration for the elasticsearch curator cronjob.

Worker

Allows to configure the workers which can be deployed along the Zeebe Cluster. The worker code can be found here.

Section Parameter Description Default
worker Configuration for the to be deployed worker application
replicas Defines how many replicas of the application should be deployed. 3
capacity Defines how many jobs the worker should activate and work on. 60

Starter

Allows to configure the starter which can be deployed along the Zeebe Cluster. The start code can be found here.

Section Parameter Description Default
starter Configuration for the to be deployed starter application
replicas Defines how many replicas of the application should be deployed. 1
rate Defines with which rate process instances should be created by the starter. 150

Publisher

Allows to configure the publisher which can be deployed along the Zeebe Cluster. The start code can be found here.

Section Parameter Description Default
publisher Configuration for the to be deployed publisher application
replicas Defines how many replicas of the application should be deployed. 0
rate Defines with which rate message should be published. 25

Timer

Allows to configure the timer which can be deployed along the Zeebe Cluster. The start code can be found here.

Section Parameter Description Default
timer Configuration for the to be deployed starter application
replicas Defines how many replicas of the application should be deployed. 0
rate Defines with which rate process instances with timers should be created. 25

Leader Balancing

Allows to configure the auto leader balancing. For more details see Rebalancing docs.

Section Parameter Description Default
leaderBalancing Configuration for auto rebalancing feature
enabled If true, enables the auto leader rebalancing. true
schedule Defines the schedule of the auto leader rebalancing feature. "*/15 * * * *"

Zeebe

For more information about Zeebe, visit Zeebe Overview.

Section Parameter Description Default
zeebe Configuration to configure Zeebe and Gateway
config Can be used to configure Zeebe Broker and Gateway additional without the need of overwriting all environment variables from the dependency chart. {}
profiling Configuration for pyroscope profiling.
profiling.enabled If true, enables the pyroscope profiling. false

Development

For development purposes, you might want to deploy and test the charts without creating a new helm chart release. To do this you can run the following:

 helm install YOUR_RELEASE_NAME --atomic --debug ./charts/camunda-platform

  • --atomic if set, the installation process deletes the installation on failure. The --wait flag will be set automatically if --atomic is used

  • --debug enable verbose output

To generate the resources/manifests without really installing them, you can use:

  • --dry-run simulate an install

If you see errors like:

Error: found in Chart.yaml, but missing in charts/ directory: elasticsearch

Then you need to download the dependencies first.

Run the following to add resolve the dependencies:

make helm.repos-add

After this, you can run: make helm.dependency-update, which will update and download the dependencies for all charts.

The execution should look like this:

$ make helm.dependency-update
helm dependency update charts/camunda-platform
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "camunda-platform" chart repository
...Successfully got an update from the "elastic" chart repository
...Successfully got an update from the "bitnami" chart repository
Update Complete. ⎈Happy Helming!⎈
Saving 6 charts
Dependency zeebe did not declare a repository. Assuming it exists in the charts directory
Dependency zeebe-gateway did not declare a repository. Assuming it exists in the charts directory
Dependency operate did not declare a repository. Assuming it exists in the charts directory
Dependency tasklist did not declare a repository. Assuming it exists in the charts directory
Dependency identity did not declare a repository. Assuming it exists in the charts directory
Downloading elasticsearch from repo https://helm.elastic.co
Deleting outdated charts
helm dependency update charts/camunda-platform/charts/identity
Hang tight while we grab the latest from your chart repositories...
...Successfully got an update from the "camunda-platform" chart repository
...Successfully got an update from the "elastic" chart repository
...Successfully got an update from the "bitnami" chart repository
Update Complete. ⎈Happy Helming!⎈
Saving 2 charts
Downloading keycloak from repo https://charts.bitnami.com/bitnami
Downloading common from repo https://charts.bitnami.com/bitnami

Releasing the Charts

Please see the corresponding release guide to find out how to release the chart.