add missing backups stale pager playbook

 - consider bumping the `max_connections` setting (in

   `postgresql.conf`) if this is a long term trend

### Backups stale

The `PgLegacyBackupsStale` alert looks like this:

    PostgreSQL backup checks are stale on test.torproject.org

This implies the backup checks have not ran recently enough. This

implies that we might not have accurate information on our backups and

should be fixed.

Those metrics are exported by

`/usr/lib/nagios/plugins/dsa-check-backuppg` script which is called

from the `/etc/cron.d/tor-backup-postgres` cron job, deployed by

Puppet. It writes to

`/var/lib/prometheus/node-exporter/tpa_backuppg.prom` which gets

collected by the node exporter "text file collector".

Check that `/var/lib/prometheus/node-exporter/tpa_backuppg.prom` has

the right timestamp, for example compare this:

    tpa_backuppg_last_check_timestamp_seconds 1728495929.425171

With the current timestamp:

```

root@bungei:~# date +%s

1728511576

```

In this case, we have a 15647 lag, which is more than 4 hours.

You should be able to run the check by hand with the command:

```

sudo -u torbackup /usr/lib/nagios/plugins/dsa-check-backuppg  -e

```

It shouldn't return any output, but it *should* update the above

timestamp.

In this case, the schedule in the cron job was incorrect. It was not

found because Nagios was running the check through NRPE and hiding the

problem, the schedule was changed to run every 15 minutes instead.

## Disaster recovery

If a PostgreSQL server is destroyed completely or in part, we need to