heph icon indicating copy to clipboard operation
heph copied to clipboard

Published 20 hours ago • verified publisher icon Thomasdezeeuw

Tracking issue for monitoring/metrics

Open Thomasdezeeuw opened this issue 5 years ago • 5 comments

Since Heph is intended for building large, distributed services, having good insights into the processes is key. Various types, such as ActorSystem, should provide data to give insights as to what is happening in the system.

I'm currently unsure on how to expose this data.

Related #13.

Thomasdezeeuw avatar Jul 10 '19 12:07 Thomasdezeeuw

Also related #31.

Thomasdezeeuw avatar Nov 26 '20 19:11 Thomasdezeeuw

Pr #330 collects some metrics, but it's closed.

Thomasdezeeuw avatar Jan 12 '21 14:01 Thomasdezeeuw

Also see https://github.com/Thomasdezeeuw/heph/issues/31#issuecomment-734459185.

Thomasdezeeuw avatar Mar 17 '21 15:03 Thomasdezeeuw

PR #462 implements some metrics, but they're just dumped out currently.

Thomasdezeeuw avatar May 15 '21 16:05 Thomasdezeeuw

Some more metrics I want to monitor:

Memory usage:

  • Heap size
  • Usage per worker/coordinator.

Per actor:

  • name
  • supervisor (?)
  • inbox/message queue size
  • priority
  • state: { Runnable, Paused } // NOTE: currently running are excluded.
  • uptime: Duration
  • polled/ran: usize
  • restarts: usize
  • time last run: Instant/SystemTime.

Maybe group the actors by name? Same for Futures?

Thomasdezeeuw avatar May 24 '21 15:05 Thomasdezeeuw