With self-hosted Linux docker runners, and especially in parallel step runs, we intermittently see failures of type `System error` with a message like:
Status 404: {"message":"No such container: c817c605-b325-5f15-a8f6-0955a544a654_52c772c6-e20b-44fe-a125-1a4226c63949_yml_docker"} The common theme is the "No such container" and we've seen many other users mentioning problems like this:
https://community.atlassian.com/t5/Bitbucket-questions/Intermittent-container-not-found-error-in-bitbucket-runner/qaq-p/2273637
https://community.atlassian.com/t5/Bitbucket-Pipelines-Runner/Runners-Autoscaler-Status-404-No-such-container/qaq-p/2659501
https://community.atlassian.com/t5/Bitbucket-questions/Error-Status-404-quot-message-quot-quot-No-such-container-system/qaq-p/2153490
My suspicion is that there is some sort of race condition happening. Either the runner is checking for the container before it's created, or perhaps more likely, after it has been destroyed.
Can an engineer on the Bitbucket end please look at the error handling code and see if there are places where querying for the container should either be retried, or a deleted container should be tolerated? Thanks.
