Closed m1kola closed 1 week ago
Name | Link |
---|---|
Latest commit | cd4084bed9d8be27b1a7ad4c2ff0aaa4d91d737c |
Latest deploy log | https://app.netlify.com/sites/olmv1/deploys/6673f1e00f968500084ec927 |
Deploy Preview | https://deploy-preview-951--olmv1.netlify.app |
Preview on mobile | Toggle QR Code...Use your smartphone camera to open QR code link. |
To edit notification comments on pull requests, go to your Netlify site configuration.
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 80.14%. Comparing base (
bfd4142
) to head (cd4084b
). Report is 4 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
It looks like in E2E we sometimes are hitting this condition and sometimes we don't which results in coverage fluctuation
Adding more test coverage is good, but it seems pretty bad that other e2e tests are non-deterministic in this way. Did you dive into this enough to understand why the flakiness is happening?
@joelanford Here is what I think happens:
I think error happens because in tests we create catalogs and we do not wait for them to become ready: we create ClusterExtension
immediately after creating a ClusterCatalog
. So most likely ClusterExtension
controller attempts to reconcile while catalog is still unpacking/not ready.
We can look into adding waits after creating ClusterCatalog
, but IMO it is good that we do not have these artificial waits: we should be able to reconcile once catalog becomes ready.
+100000 to not artificially wait. Expecting eventual consistency in our e2e's is the correct approach.
This makes sense. Thanks for the explanation.
Description
Cover what happens when we fail to fetch bundles from catalog.
This should solve flaky test coverage. It looks like in E2E we sometimes are hitting this condition and sometimes we don't which results in coverage fluctuation:
Reviewer Checklist