Open jack-berg opened 1 year ago
Interesting. I imagine these would be sent as log events and would be opt-in, right? Would the detection be always on or could it be triggered by something else? I wonder how frequent these events are...
Good thing we're just renaming runtime-metrics
to runtime-telemetry
😄
I think this could be a nice addition; could be disabled by default - we're not exporting logs by default anyway (at least for now).
agine these would be sent as log events and would be opt-in, right?
Minimally opt in for a while. Could imagine later turning on by default if they get added to semantic conventions.
Would the detection be always on or could it be triggered by something else?
I figure it could just run periodically with a interval long enough that there isn't concern about execution time.
I wonder how frequent these events are...
I hope infrequent! 😁
I'm curious if there's an appetite for adding JMX based deadlock detection. I think deadlocks are good candidates to emit as events because:
Here's a quick and dirty prototype that detects deadlocks and emits as events:
If I cause a deadlock on an app configured with the
OtlpJsonLoggingLogRecordExporter
, I get data shaped like:It would be nice to be able to detect and alert on deadlock events.