Auto-retry failed builds




Sometimes builds can fail due to random issues like network timeout, random spec failure etc.

It would be great if circle.yml config allowed to specify retry option with an integer value that would mean number of automatic retries before circle treats this build as a failed one.

  retry: 2

That would allow to include this setting in particular workflows only. For example we want to be able to retry specs only in staging and production workflows, but not in feature branches, where the chance if non-random spec failures is pretty high

Thank you for consideration.