[ENHANCEMENT] Extend QA Values Files For Both Tasklist V1 & V2

[slolatte]

Describe the use case:

Right now the QA values files use the default Tasklist mode (V2). However, to extend the test coverage, we would like the option to additionally set the Tasklist mode to V1 and run our automated tests against this as well.

Describe the enhancement/feature:

This feature involves updating the reusable Helm chart setup workflows to support the Tasklist mode as an input. The relevant workflows to be updated are:

Helm Chart Setup (8.8 Plus)

Desired outcome and acceptance tests:

[hisImminence]

This config can be set via: https://docs.camunda.io/docs/components/tasklist/api-versions/#self-managed-configuration

CAMUNDA_TASKLIST_V2_MODE_ENABLED=false

[slolatte]

Something to keep in mind when implementing this: for each test run (across all scenarios), we’ll need to execute some tests in Tasklist V1 Mode and others in Tasklist V2 Mode.

With our current setup, that would mean spinning up two clusters per workflow. My concern is that once the Always Green Epic is in place, the number of clusters running on the Distro CI environment will already increase significantly - and this approach would effectively double that load.

It might be a good time to explore whether we can dynamically update the cluster configuration instead. For example:

• Spin up a cluster in V1 Mode → run V1 tests
• Update it to V2 Mode → run V2 tests

This could help optimize resource usage and keep CI runs efficient.

[hisImminence]

Right, this would be related to #2770 - right?

I suggest to keep the scope of this ticket as a first iteration and provide a configuration only to enable V1 - then lets iterate on the other ticket about the question on how to scale this setup..., sounds good?

[slolatte]

@hisImminence Yes and sounds like a plan!

[hilalbursalii]

Hey @eamonnmoloney
I'm applying some changes for this ticket on our side. I was wondering do you have an estimated timeline for when this ticket can be picked up or implemented?

[eamonnmoloney]

Hi. What is the urgency attached to this ticket? There is have some refactoring I'd like to do on the helm repo before added this scenario.

Also, it is mentioned that you'd like to deploy for V2 first and then switch to V1. Is this a hard requirement? Unless this is expected customer behaviour (i.e. to change the API version on a running cluster) I'd rather stick to a single greenfield install for change. A change like this to a running cluster is effectively the same as an upgrade scenario.

[hilalbursalii]

@eamonnmoloney We don’t have any coverage for v1 at the moment, since the default is v2. I think it’s important to complete this before the next alpha. I’ll also need to apply some related changes after this ticket, so I’ll need a bit of time for that as well.

Regarding the dynamic configuration, I believe it’s related to the increased load when specifying two HCs per flow, since we want to be able to run tests in both v1 and v2 modes for versions ≥ 8.8.

@slolatte Could you confirm if that’s the right priority and if my understanding is correct?

[slolatte]

Thanks both!

For 8.8 and 8.9 clusters, we’ll need to test against both v1 and v2 (by 8.10 it will just be v2). Since the default is already v2, it’s unlikely customers would go from v2 → v1. However, what is expected is the v1 → v2 migration path.

In that case, the test flow would ideally be:

1. Deploy a cluster using v1
2. Run our v1 tests
3. Update the cluster to v2
4. Run our v2 tests

As @hilalbursalii mentioned, we currently don’t have automated HC coverage for v1, so it’s important to have this in place for 8.9.0-alpha1 to ensure proper integration.

@eamonnmoloney To stay aligned with our release testing plan, we’d need the related Distro work completed by next Friday (Oct 31). Do you think that timeline is doable given the refactoring you mentioned?

[eamonnmoloney]

@slolatte, this is 3rd on my list of things to get done..

• first is medic (quite so far so hopefully it stays that way)
• second, change the infra on CI so it is using external ES and KC. This will speed up everything and greatly help resource issues on the CI
• third is the tasklist v1 and v2 (so 2 new workflows)
  • this will be install just v1
  • install v2 already done
  • install v1 (8.7) and upgrade to v2 (8.8)

As long as medic doesn't become a big thing, this will be done. I'll message you if I think it is at risk