Refactor to ignore a failure for the grace period

[gregcorbett]

Supersedes #18, depends on #23 and #22. Resolves GT-235.

• currently, an error in the pipeline immediately causes the next icinga check to raise an error.
• this led to flapping hour to hour, generating emails that were ignored.
• this should mean we only get alerts if the failover pipeline has not succeeded within the grace period.
• adding an output on a successful restore in 1_runDbUpdate.sh simplifies the logic needed by this check. We don't have to worry about "if/when the pipeline failed?", we only care about "when did it last succeed?".

This PR will also add:

• unit tests for the check_db_dump_recent.py script
• a GitHub workflow to run those tests at PR time
• a dependabot config to keep workflows/actions up to date

[gregcorbett]

The new unit tests do pass: https://github.com/gregcorbett/gocdb-failover-scripts/actions/runs/6036355188

[tofu-rocketry]

Could you add some notes to the PR about the GitHub workflow changes? (Assuming they are intentionally part of this PR.)

[gregcorbett]

Done

[gregcorbett]

I've squashed the commits down a bit, but GitHub confirms no changes have snuck in.

[tofu-rocketry]

We're all green then. :shipit: