We have monitored our top ci failures in our system,

Our Jenkins is running on top of AKS with dynamic workers we saw that the top common issue related to 'retrieveChanges' mostly a 'clone' operation re-running the job solves the problem of course we added a simple retry.

In the initial clone (fetch Jenkinsfile) there is already a re-try mechanism but in the 'retrieveChanges' part, there is no retry.

We have like 1000 jobs running in parallel and it doesn't make sense to implement retry mechanism in all of these Jenkins files.