CODEWITHSUNDEEP
datadog
Zoinks
Zoinks

Reputation: 361

How can I alert that a particular process is crashing in Datadog?

I'm trying to figure out how to create an alert around a process that may be crashing and restarting repeatedly. It might be providing some data to Datadog while it's up, so a "no data" alert won't do because the lack of data never hits the duration threshold as the process restarts. I was thinking of alerting on a changing PID, but I cannot for the life of me figure out how to create a PID-based Monitor. Is it possible? And how? Does anyone have any other suggestions for this situation?

Upvotes: 0

Views: 740

Answers (1)

stephenlechner
stephenlechner

Reputation: 2269

Maybe you could mod the process check to also tag the process number metric by PID (this is probly where you'd change that). That way you could group your monitor by your pid tag and the no-data alerts would tell you when the pid switched.

But this would also alert on expected pid changes, so maybe you'd have to schedule downtimes too aggressively for this to be a good idea?

Maybe monitoring some crash logs with their Log Management tool would be a better approach?

Upvotes: 1

Related Questions