kacpermuda commented on code in PR #69671:
URL: https://github.com/apache/airflow/pull/69671#discussion_r3557983463
##########
.github/workflows/additional-prod-image-tests.yml:
##########
@@ -329,34 +325,17 @@ jobs:
e2e_test_mode: "go_sdk"
if: inputs.canary-run == 'true' || inputs.run-go-sdk-e2e-tests == 'true'
- test-providers-e2e-tests-openlineage:
- name: "Provider E2E tests openlineage"
- uses: ./.github/workflows/providers-e2e-tests.yml
- with:
- provider: "openlineage"
- provider-display-name: "OpenLineage"
- runners: ${{ inputs.runners }}
- platform: ${{ inputs.platform }}
- default-python-version: "${{ inputs.default-python-version }}"
- use-uv: ${{ inputs.use-uv }}
- if: inputs.canary-run == 'true' ||
inputs.run-providers-e2e-tests-openlineage == 'true'
-
- test-providers-e2e-tests-openlineage-compat:
Review Comment:
I think that it's valuable to keep all the versions in compat, as OL is one
these "meta" providers that has a lot of reliability on core, so it's easy to
break some compatibility by something that looks like small and innocent
change. These compat jobs are already using airflow images published, just
adding some providers on top of it, so they are not very heavy. I do however
understand the need to save the CI resources, I think if it turns out that it
is consuming too much, we can always trigger these compat jobs only for
scheduled runs on main (those 4 times a day if I remember correctly?) And only
run them with prod image for PRs, that way, we can reduce the resource usage
and still keep the early bug detection.
How does this sound?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
