Monitor and/or remove unused SSD's

Created by: caugustus-sourcegraph

Followup to https://github.com/sourcegraph/sourcegraph/issues/31262: we still see unused SSD's left behind in some GCP projects. These disks can be leftover from failures during automated testing, incorrectly deleted deployments, or manual tests performed in testing projects.

We'd like to have some kind of notice when disks have been abandoned for over 24 hours so they can be investigated and removed. We can also consider auto-removing the disks (maybe only in select projects) to remove any manual intervention required.