Monitoring is supposed to observe, not add load
A monitoring check that hits a site too frequently, or too heavily, can meaningfully contribute to the very server pressure it exists to detect — especially on a smaller server already running several sites.
We've seen this happen on a real client server: a genuine memory-pressure incident turned out to be partly caused by our own monitoring checks running more aggressively than the server could comfortably absorb.
The fix is pacing, not less visibility
The answer isn't to monitor less — it's to monitor in a way that respects the resource it's watching: sensible check intervals, lightweight requests, and awareness of what else is running on the same server.
Good monitoring should tell you about a real problem before your users do, without ever being the reason there's a problem to tell you about.