MySQL acquisitariat server (for DevCentral and other Equatower services) is sometimes heavily solicited.

Some monitoring would be welcome there.

MySQL Workbench produces this kind of dashboard for example:

More urgently, I would like we sometimes check SELECT count(*) FROM phabricator_daemon.daemon_logevent. I truncated the table this night, it was at 1 519 666 rows rows.

Phabricator is happy to fire a lot of daemons error quietly to MySQL and /var/tmp/phd/log/daemons.log without any reporting on the UI. To detect high count for this table would be a good indicator something is wrong.