Verify bridge jobs receive status updates from downstream pipelines
As we migrate more deployer functionality into release-tools, we're relying on knowing the status of the downstream deployer pipeline.
Currently we trigger it, and then a delayed job checks its status at some point in the future. This has some downsides outlined elsewhere, and one possible fix is to use bridge jobs.
This came with a caveat that if the pipeline triggered by the bridge job failed, but then succeeded on retry, the bridge job didn't receive an updated success
status, thus preventing future stages in the parent pipeline from running.
This was recently fixed and deployed, so we should verify if it will work for our coordinated pipelines.
Testing
- Create two test projects,
Parent
andDownstream
. - Enable
ci_reset_bridge_with_subsequent_jobs
feature flag in both projects. -
Parent
contains a bridge job that triggersDownstream
, and a job in a subsequent stage. -
Downstream
contains a job that fails on the first run, and then passes on a retry. We can probably just do this with CI variables. - Verify that
- The bridge job status changes from
failed
tosuccess
after the retry - The subsequent stage runs after the status change.
- The bridge job status changes from
Edited by Robert Speicher