search

Altinn / altinn-notifications

Altinn platform microservice for handling notifications (mail, sms, etc)
MIT License
2 stars 3 forks source link

Analysis workflow reports failing tests when all tests pass #594

Closed acn-sbuad closed 3 weeks ago

acn-sbuad commented 3 weeks ago

Description of the bug

see title.

Issue created in repo of action creator NasAmin/trx-parser#311 Example workflow run: https://github.com/Altinn/altinn-notifications/actions/runs/10453675204/job/28944613679

Steps To Reproduce

Currently disabled tasks in workflow: https://github.com/Altinn/altinn-notifications/blob/main/.github/workflows/build-and-analyze.yml Remove comments and trigger workflow to reproduce error.

SandGrainOne commented 3 weeks ago

Are we able to download the trx files to review them?

In some of the cases I've seen there have been errors after the tests have been run: The active test run was aborted. Reason: Test host process crashed : Unhandled exception. Npgsql.PostgresException (0x80004005): 53300: remaining connection slots are reserved for non-replication superuser connections

It could be that it does something to the report files.

acn-sbuad commented 3 weeks ago

@SandGrainOne We would need to adapt the workflow to be able to download the reports.

In the case you added to the example the Build&Test step actually fails and does not exit with a success code. This is not the same as the case. Removing the additional information, to make the issue clear.

In the current case, there are no logs indicating that anything has failed.

acn-sbuad commented 3 weeks ago

Could not reproduce as of now. Made test report available as an artifact so we can check closer if it is to occur again.

acn-sbuad commented 3 weeks ago

Test reports show that an error message being logged when connecting to the kafka broker is causing the test run to be determined as failed. The error message Failed to acquire idempotence PID from broker localhost:9092/1: Broker: Coordinator load in progress: retrying. It is reasonable to assume that retry is successful as all test that depend on kafka are successful. Question is how we avoid the error message or ensure it does not break the final result of the test run.

Testrun file is available as an artifact on the workflow