Revamp CI

[chrisseto]

The CI setup (in both helm-charts and the operator) are a bit all over the place right now. The recent introduce of go code into the helm-chart adds in a large chunk of untested code and now generated files that can easily get out of date. We need to comb thought the existing configurations and collapse them down into a single definition that's usable for CI and local development a like. We'll also want to rely on github actions for helping us maintain our versioning policies (#1093 ).

Tasks

Beta Give feedback

1. Consolidate on Taskfiles for "step" definitions #1105
   +0
2. Consolidate on nix for environment/dependency management #1106
   +0
3. Settle on a naming convention for taskfiles
4. Collapse all Github Actions for pull requests into a single workflow #1104
   1 of 1
   +0
   
5. Ensure that nightly tests are running for supported versions of Kubernetes, Redpanda, the operator, and helm charts #1254
   +0
6. Ensure we have github actions for bumping various versions automatically and that they commit as a bot user or at least someone currently employed at redpanda.
7. Utilize Github Actions for automating releases (Push a tag and have a release minted, it may be good to make some branching conventions)
8. Chart-testing github action is flakey or broken #1097
   +0
9. Lint generated files (README.md, values.schema.json, gotohelm files, etc) #1103
   +0
10. Version bump actions don't update golden tests #1109
    +0
    
11. Generate all of values.yaml, values.schema.json, and values.go from a single source of truth #1081
    2 of 2
    +0
    

[chrisseto]

Going to close this out and move the unfinished tickets into the backlog. We're in a pretty good place for the helm-charts and an okay place for the operator.