ci: add Helm chart validation

[avirajkhare00]

Summary

• add a helm component to change detection so helm/** routes through the existing other_matrix
• add pinned Helm install, helm lint --strict, and the 6 requested render scenarios in _test.yml
• make the legacy ingress and HPA API selection deterministic by Kubernetes version so the 1.18 CI case really covers networking.k8s.io/v1beta1 and autoscaling/v2beta2

Closes #3007

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 71.74%. Comparing base (9ee1f6d) to head (5f99cfd).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #3019      +/-   ##
============================================
- Coverage     71.76%   71.74%   -0.02%     
  Complexity      943      943              
============================================
  Files          1121     1121              
  Lines         93800    93800              
  Branches      71125    71135      +10     
============================================
- Hits          67315    67298      -17     
- Misses        23851    23852       +1     
- Partials       2634     2650      +16     

Components	Coverage Δ
Rust Core	72.35% <ø> (+<0.01%)	⬆️
Java SDK	62.30% <ø> (ø)
C# SDK	67.43% <ø> (-0.21%)	⬇️
Python SDK	81.43% <ø> (ø)
Node SDK	91.44% <ø> (-0.10%)	⬇️
Go SDK	38.68% <ø> (ø)
see 22 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[avirajkhare00]

I dug through the hardcoded bits in this PR. The main problem is not the number of scenarios, it is that _test.yml is now carrying some chart-owned values that already exist under helm/charts/iggy/.

There are two buckets here:

1. Reasonable CI-owned hardcoding

• pinned Helm/kind versions and checksums
• kind cluster config
• ingress-nginx install
• smoke-test hosts / ingress class / namespace / release names

Those are properties of the CI environment and smoke path, so keeping them in the workflow is fine.

2. Chart-owned values that should ideally be reused from helm/charts/iggy/

• chart version / appVersion from helm/charts/iggy/Chart.yaml
• default image tags from helm/charts/iggy/values.yaml
• render scenario inputs that are really chart fixtures rather than CI wiring

That second bucket is where the PR feels too hardcoded, because it creates a second source of truth in CI.

If you want to keep this structure, I would suggest a cleaner split:

• leave CI/runtime setup in _test.yml
• move chart-specific test fixtures into the helm/charts/iggy/ tree (for example dedicated values files under a small ci/ or tests/ subdir), or derive the asserted values from chart files directly

That keeps the Helm chart as the owner of chart data and makes version/tag bumps less noisy.

[hubcio]

overall looks nice, but please don't put bash scripts in _test.yml - please move the actual scenario (not setup) to scripts/ci/test_helm.sh and call it from action context. we want to allow the devs to run this locally. it has to work this way too.

[avirajkhare00]

@hubcio pretty sure c# issues are not happening from helm tests.

Edit, added a small fix here: #3024

Also, I have ran all the smoke tests of helm manually. so both ci and humans can run it

[hubcio]

also please rename test_helm.sh to test-helm.sh for consistency with the other ci scripts (setup-helm-tools.sh, setup-helm-smoke-cluster.sh all use hyphens).

[hubcio]

@avirajkhare00 looks like your newly added smoke test failed

[avirajkhare00]

@hubcio race condition it was: https://stackoverflow.com/questions/68576225/failed-calling-webhook-validate-nginx-ingress-kubernetes-io-error-while-apply

thankfullly its fixed