Closed yama6a closed 1 month ago
Thanks @yama6a - for e2e testing, see our tests in https://github.com/GoogleCloudPlatform/prometheus-engine/tree/main/e2e.
They deploy a full stack running to a kind cluster and do assertions on the workloads therein. You could add some testing there to double-check the manifests etc
Thanks @yama6a - for e2e testing, see our tests in
main
/e2e.They deploy a full stack running to a kind cluster and do assertions on the workloads therein. You could add some testing there to double-check the manifests etc
@pintohutch Thanks for the pointer. I'll see what I can do 🙂
Hi @bwplotka! Im Henrik, colleague of Yama here. We have added to the e2e tests as requested, please have another look at our PR.
@bwplotka @pintohutch
We added an e2e test. We also deployed the changes to the rule-evaluator, the k8s service, and the frontend to a GKE test-cluster and smoke-tested the endpoints. Everything works as expected: rule-evaluator endpoints, the service, and the endpoint-forwarding: frontend-pod -> rule-evaluator-service -> rule-evaluator-pod
We refactored quite a lot in this PR compared to the first commit, improving on the structure and testability of the code, so you might want to re-review.
Nice! I had some readability comments... ended up doing a PR to fix this, might be easier. I left the functionality untouched really -- great job!
My recommendations are here yama6a#1
Thank you! Your suggestions have been applied, ready for another round.
Ive squashed the commits to comply with the presubmit check for commit message format
Thanks! 💪🏽
Enabling to query
/api/v1/rules
and/api/v1/alerts
from thefrontend
by proxying these calls to therule-evaluator
pod.This PR also creates a k8s service manifest for the
rule-evaluator
in order to be able to reach it fromfrontend
Note:
There was no proper way to test the k8s manifests in an actual cluster. If there is a way to test this setup properly before merging, that would probably be advisable.Update: Added e2e tests