I noticed from #150 and some PRs (#175) that maintaining the chart and having to manually test the changes is quite cumbersome.
So I would propose adding Integration and E2E testing with kind to reduce the amount of manual work needed and verify that changes do not have unintended side effects.
I wanted to create this issue first before creating yet another (big) PR to discuss if and what is worthwile.
Currently I came up with a small setup that just (matrice) tests the current charts and runs the helm tests that are already shipped to smoke test the deployment. See: https://github.com/valkey-io/valkey-helm/compare/main...DanielHabenichtWork:valkey-helm:valkey-ci-deployment-tests?expand=1
Later on the functionality and failure recoveries could be tested with Chainsaw: https://github.com/valkey-io/valkey-helm/compare/main...DanielHabenichtWork:valkey-helm:valkey-ci-deployment-tests-chainsaw?expand=1
I noticed from #150 and some PRs (#175) that maintaining the chart and having to manually test the changes is quite cumbersome.
So I would propose adding Integration and E2E testing with kind to reduce the amount of manual work needed and verify that changes do not have unintended side effects.
I wanted to create this issue first before creating yet another (big) PR to discuss if and what is worthwile.
Currently I came up with a small setup that just (matrice) tests the current charts and runs the helm tests that are already shipped to smoke test the deployment. See: https://github.com/valkey-io/valkey-helm/compare/main...DanielHabenichtWork:valkey-helm:valkey-ci-deployment-tests?expand=1
Later on the functionality and failure recoveries could be tested with Chainsaw: https://github.com/valkey-io/valkey-helm/compare/main...DanielHabenichtWork:valkey-helm:valkey-ci-deployment-tests-chainsaw?expand=1