testkube regression: helm chart fails to install with flux due to duplicate yaml keys

[gberche-orange]

Describe the bug

#449 introduced a regression with duplicate yaml key between the template and default values

helm-charts/charts/testkube-api/templates/deployment.yaml

Lines 178 to 183 in 68dd5e2

	readinessProbe:
	{{- toYaml .Values.readinessProbe | nindent 12 }}
	httpGet:
	path: /health
	port: {{ .Values.service.port }}
	resources:

helm-charts/charts/testkube-api/values.yaml

Lines 314 to 318 in 68dd5e2

	## Testkube API Liveness probe
	livenessProbe:
	httpGet:
	path: /health
	scheme: HTTP

As a result, installation from fluxcd fails with message

      Helm upgrade failed: error while running post render on files: map[string[]interface {}(nil): yaml: unmarshal errors:
        line 107: mapping key "httpGet" already defined at line 100
        line 118: mapping key "httpGet" already defined at line 112

See related fluxcd/helm-controller#283 (comment)

While duplicate keys are mostly ignored by some tools but not others, this problem is unfortunately not detected by helm lint (once #452 gets included)

$ helm lint
==> Linting .
[INFO] Chart.yaml: icon is recommended

1 chart(s) linted, 0 chart(s) failed

More background explaining why fluxcd fails with duplicated yaml keys can be found at kubernetes-sigs/kustomize#3480

I searched a while how to automatically detect such regressions in the future in the kubeshop/helm-charts github workflows:

• some python libraries can detect duplicated yaml keys
• the sigs.k8s.io/kustomize/kyaml cli is likely to detect such errors

https://github.com/kubernetes-sigs/kustomize/blob/2fa944b1cdc1f9ad4f72b3df8e67553e0b97d37d/cmd/kyaml/README.md

This package exists to expose kyaml filters directly as cli commands for the purposes of development of the kyaml package and as a reference implementation for using the libraries.

I for now failed to install kyaml (likely by lack of skills related go build environment)

go install -v sigs.k8s.io/kustomize/cmd/kyaml@latest
go: sigs.k8s.io/kustomize/cmd/kyaml@latest: module sigs.k8s.io/kustomize@latest found (v2.0.3+incompatible), but does not contain package sigs.k8s.io/kustomize/cmd/kyaml

To Reproduce

Steps to reproduce the behavior:

helm install --dry-run testkube-manual kubeshop/testkube --version 1.10.272  | tail -n +7 ../testkube.yaml | head -n -3 > testkube.yaml

Find duplicate keys

          readinessProbe:
            failureThreshold: 3
            httpGet:
              path: /health
            initialDelaySeconds: 30
            periodSeconds: 15
            successThreshold: 1
            timeoutSeconds: 10
            httpGet:
              path: /health
              port: 8088

** workaround **

Revert to 1.10.256

Expected behavior
A clear and concise description of what you expected to happen.

Version / Cluster

• Which testkube version? helm chart 1.10.257 to 1.10.272
• What Kubernetes cluster? (e.g. GKE, EKS, Openshift etc, local KinD, local Minikube)
• What Kubernetes version?

Screenshots
If applicable, add CLI commands/output to help explain your problem.

Additional context
Add any other context about the problem here.

[ypoplavs]

This issue was fixed in our new release version. Please upgrade to the latest version and let us know in case of anything.

[gberche-orange]

thanks @ypoplavs for the fix in a21af64 ! This indeed solves the reported issue.