Parameterize PodDisruptionBudget config for Linkerd Control Plane components

[cdalar]

Parameterize PodDisruptionBudget config for Linkerd Control Plane components

The maxUnavailable setting for all Linkerd Control Plane components is hardcoded to 1

Parameterize PodDisruptionBudget config in templates, which is currently hardcoded (maxUnavailable), so that it can be configured from values.

Fixes #11321

Signed-off-by: Cemal Y. Dalar [email protected]

[cdalar]

@alpeb , did the changes you mentioned. Hope that is good.

[alpeb]

A unit test in values_test.go is failing, as you can see here: https://github.com/linkerd/linkerd2/actions/runs/7297741456/job/19896871341?pr=11687#step:6:172 You need to edit that file and fill in an appropriate value in the expected Values struct for PodDisruptionBudget.MaxUnavailable which now is different for the HA case.

[cdalar]

@alpeb . Do we need to ask from someone with write access to approve as well? @olix0r or do you need something else from me?

[kflynn]

@cdalar We got a couple of folks to look over this, and realized that we'd like one change – could you move podDisruptionBudget under controller, so that you'd set controller.podDisruptionBudget.maxUnavailable? (We're trying to be better about namespacing new parameters...)

Thanks! 🙂

[cdalar]

@cdalar We got a couple of folks to look over this, and realized that we'd like one change – could you move podDisruptionBudget under controller, so that you'd set controller.podDisruptionBudget.maxUnavailable? (We're trying to be better about namespacing new parameters...)

Thanks! 🙂

@kflynn . 👍 related changes have been made. Please check..

About namespacing; there are other parameters which should be inside the new controller object but I think it's out of this scope. Maybe later I can tackle them as well.

[cdalar]

@alpeb can you please re-run the tests..

[kflynn]

Hey @cdalar – I reran the test just to check, but it looks like a real failure. As @alpeb noted above, this one will need some test data to be regenerated:

• go test ./... -update should update the test fixtures, then
• ./bin/helm-docs should re-generate the README.md file with the new values.yaml entry.

Try that? and then let's see where we are. 🙂

[cdalar]

Hey @cdalar – I reran the test just to check, but it looks like a real failure. As @alpeb noted above, this one will need some test data to be regenerated:

• go test ./... -update should update the test fixtures, then
• ./bin/helm-docs should re-generate the README.md file with the new values.yaml entry.

Try that? and then let's see where we are. 🙂

@kflynn Run both commands. The install_output.golden has been updated now. I hope that is enough.