gitlab-restore: implement cleanup procedure to avoid "out of quota" events
It's quite common now the cases when gitlab-restore project reaches some quotas because some instances are stalled. If, during restoration, some error occurs, in some cases it doesn't lead to hard failure and auto-cleanup.
We need some job that would clean up instances periodically, based on certain mask for instance name and rules like "If the backup verification started 3 days ago and hasn't finished yet, and if the instance is not protected from deletion, it's time to consider backup verification as failed, send all the signals about it, and destroy the instance".
To discuss: how should it be organized? Cronjob or CI/CD pipelines with special task? Anyhow else?