I don't track their performance, I just track if they're up or down.

I use uptimekuma running on a free tier of fly.io so I can tell if my cluster had a catastrophic failure. There's no point in the alerting system running on the same system.

If you do truenas to truenas you could schedule a zfs send.