-
Notifications
You must be signed in to change notification settings - Fork 148
Description
This epic summarizes the remaining work for our test setup required for the QA team.
Over the last weeks we made tremendous efforts on improving the test coverage and setup 👏 and thanks @eamonnmoloney. The configuration of infrastructure (HC configuration as values.yamls and the workflow setup) is now mostly in the hands of the Distribution Team.
Current state
- Over the past weeks we faced still some issues (e.g. too long names for namespaces, resource issue for HC setups) which lead to high instability:
- By increasing the resources it seems this issue is fixed (still needs confirmation/ proof)
- Also the names of the namespaces have been shortened.
- TTL was set down. QA was asking if it could be back on 4 hrs. in the resource issue is resolved.
- Its possible to set the TTL when kicking off a workflow. Per default, the TTL is 1hr.
- See [ISSUE] Helm Chart Setup Failing Consistently Across SM Workflows — Blocking Nightly Tests and On-Demand flows #3920
- @aabouzaid to confirm issue is resolved and TTL can be increased to 4 hrs. without having again resource issues.
- We have still remaining PRs in progress for test scenarios and setup improvements -it should all be done until the 8.8 release (latest mid septmeber).
- zeebe CPU was increased to 4 cores.
Open PRs/ Todos
- License Testing
- https://github.com/camunda/c8-cross-component-e2e-tests/pull/1059
- @slolatte is reviewing this PR
- fix: implement downgrade protection for the helm charts #3560
- https://github.com/camunda/c8-cross-component-e2e-tests/pull/1108 ("PR-triggered-workflow" (lower prio)
- [ISSUE] EKS Cluster HC Setup Fails with Error #3865
- Make Distro nightly tests green again: https://github.com/camunda/camunda-platform-helm/actions/workflows/test-version-nightly.yaml
- [ISSUE] issues with the introduction of 8.7 e2e tests for every PR run #3809
- https://github.com/camunda/team-distribution/issues/546
Missing Test scenarios
- S3 Bucket/ document handling:
- Values.yaml is ready to be used
- Missing workflow to trigger HC deployment + QA test
- feat: add documentStore to QA ES values #3997
- OIDC EntraID
- values.yaml is ready to be used
- Missing workflow to trigger HC deployment + QA test
In the QA-Distro sync on the 2025-08-06 it became clear that its unclear who is responsible for creating the workflow for the HC build (e.g. S3 bucket or OIDC EntraID) and it needs further discussion with @eamonnmoloney - For now its important that we get unblocked on the 2 workflows to test the S3Bucket + Microsoft EntraID.
Feel free to adjust and extend this ticket.
Outlook
- Other topics to explore after the main issues (🔝 ) are resolved are:
- Monitoring and Alerting
- Provision of new environments
- Stability of tests and workflows
- Responsibility/ Ownership
- Test suits and reusability of tests in SM (e.g. C8Run, Docker Compose, ...)
- Testing of infrastructure
We are thinking of planning a workshop around QA + SM Team after the 8.8 release to align on these topics.