Automatically close broken master incidents for known failures

Proposal

We should automatically close the broken master incidents that are related to known failure reasons:

• master-brokeninfrastructure (e.g. a runner failed to pull the docker image)
• master-brokendependency-upgrade (e.g. a segmentation fault error in #390313 (closed))

Also, we have a list of identified failure reasons: https://gitlab.com/gitlab-org/quality/engineering-productivity/team/-/blob/main/runbooks/master-broken.md#known-failure-reasons, which can be a great starting point for automating this.