Yeah, it’s basically a handler for nagios/sensu/etc which will open/close issues in its DB and assign them to whomever is on-call for that day.

It can then phone/text them when it gets an alert, and escalate to the next person in the list of it doesn’t get an answer.

It’s not super-hard to replace with some twilio scripts/etc but it’s cheap and usually works as expected.

It’s nice to pay them to run alerting scripts, so they don’t run on the same infrastructure as production code.

Events are whatever you send them. So we have some internal monitoring tools and SignalFX set to trigger pager duty when certain metric thresholds are tripped