How are you dealing with alert fatigue from your monitoring tools?

I was using an existing monitoring app to monitor ~50 websites.
And it was working, but some issues were there.

It was sending a lot of emails for some websites that are not stable due to some random 500 errors or time out errors.
I was tired with lots of email alerts for 2-3 websites I added, all other websites were stable.

Most of the time nothing was actually “down.”
It was just temporary instability.

I was worried it will cause neglecting the real down alerts.

So I built my own approach. I created a new uptime monitoring tool, with a focous.
Never send an alert without multiple checks.
Send an alert only after 4 checks are done, that too from multiple locations.

Now before we send an alert:
• We retry 3–4 times
• We verify the monitor is down from multiple locations
• We make sure it’s actually consistently failing

If it’s just a minor flip, we ignore it.
If it’s really down, you’ll know.

Way fewer false alarms.
Way more confidence when an alert does fire.

Curious — how are you handling noisy alerts in your stack?
If any one would like to check it, please see https://statuseagle.com/