CirrusCI/FreeBSD: Monthly compute limit exceeded!

[cmb69]

Description

This apparently happens often at the end of month, and I think this is bad, because we may easily have no CI feeback for FreeBSD for days (or even weeks, if the amount of pushes increases). So unless somebody would pay for a compute limit increase, maybe we should stop running CirrusCI for every push, but instead have nightly runs, and a way to manually trigger the workflow in cases where it seems relevant. I don't know if that's possible, though.

cc @devnexen

[iluuu1994]

We discussed this when Cirrus stopped their unlimited free tier. The main "issue" with testing nightly is that Cirrus doesn't have a built-in way to send e-mails on error. So, we'd have to implement that (or Slack notifications) ourselves. Since we thought just moving the ASAN build to GitHub was enough, we did that.

[cmb69]

Oh, we switched to GH actions quite a while ago, so this has already been resolved.