Bandstand Charts

This is a public repo since it uses Github Pages to host a helm repository.

Chart References

Cron Job

Triggered Job

Test Runner

Web Service

Headless Service

Job

Confluent Connect Cluster

Values precedence

Values in values-env.yaml will override values in values.yaml. The properties in the environment-specific file will be merged together, and when the same value is specified in both places the env-specific values will take precedence. The only exception is the environment variables values because they are defined as a list and helm doesn't support merging lists, that is why the charts provide two different values to for setting env vars additionalEnvVars (for values.yaml) and environmentEnvVars (for values-<env>.yaml).

NB: The precedence is actually based on the ordering in the CircleCi config (last file wins). The templates specify the env values file later, but if you modify the ordering the behaviour will change.

Updating Library Charts

• Create branch for updates.
• Update charts/{name}/templates with changes.
• Bump charts/{name}/Chart.yaml version.
• Create PR from branch.
• Merge when approved, self-hosted helm repo will be updated.

Check the available versions here.

Using Library Charts

Add as a dependency in your Chart.yaml:

dependencies:
  - name: bandstand-web-service
    version: {version}
    repository: https://ktech-org.github.io/bandstand-charts/

And update dependencies:

helm dependency update .

This will add the dependency to the chart's charts subdirectory.

Migrating to Advanced Scaling

The bandstand-web-service and bandstand-headless-service charts both support advanced scaling options using KEDA. This enables scaling to zero, scheduled scaling and scaling on external triggers such as kafka messages. To see examples of scalers that can be used see the test chart

To migrate to advanced scaling you must first disable the HPA by setting hpa.enabled to false in your values.yaml file. That change should be deployed to all clusters before proceeding. As a separate deployment you can enable advanced scaling using KEDA by setting advancedScaling.enabled to true in your values.yaml file and adding the required scalers.

If HPA is disabled and advanced scaling is enabled in the same deployment, the deployment will fail. If this is the case contact the dev platform team for assistance as we will need to manually delete the old HPA resources from the clusters to allow the deployment to proceed.

If you are interested in enabling other KEDA scalers, please contact the dev platform team or raise a PR.