update the time we consider a backup timestamp "too old" from 30 minutes to 45 minutes. As Concourse has gotten busier, the backup checks now run a more than a few minutes late, but that should be ok for now
update link to point to the actual home for the repo
I think the real long-term fix here is to move these checks out of Concourse and onto prometheus or the maintenance VM as a cron job.
security considerations
Having stale backups is bad, and this increases the chances that backups will go stale without notice
On the other hand, this alert fires false alarms so frequently that people don't look at it anyway, so improving the signal:noise ratio here is a win
Changes proposed in this pull request:
I think the real long-term fix here is to move these checks out of Concourse and onto prometheus or the maintenance VM as a cron job.
security considerations
Having stale backups is bad, and this increases the chances that backups will go stale without notice On the other hand, this alert fires false alarms so frequently that people don't look at it anyway, so improving the signal:noise ratio here is a win