-
Notifications
You must be signed in to change notification settings - Fork 557
Open
Description
opentelemetry_sdk version: 0.29
If I enable debug logging for the crate I see the following:
2025-06-09T19:45:59.696986Z DEBUG opentelemetry_sdk: name="BatchLogProcessor.ExportingDueToForceFlush"
2025-06-09T19:45:59.697019Z DEBUG opentelemetry-otlp: name="TonicsLogsClient.CallingExport"
2025-06-09T19:45:59.697580Z DEBUG opentelemetry_sdk: name="LoggerProvider.ShutdownInvokedByUser"
2025-06-09T19:45:59.697604Z DEBUG opentelemetry_sdk: name="BatchLogProcessor.ExportingDueToShutdown"
2025-06-09T19:45:59.697625Z DEBUG opentelemetry-otlp: name="TonicsLogsClient.CallingExport"
2025-06-09T19:45:59.698071Z DEBUG opentelemetry_sdk: name="BatchLogProcessor.ThreadExiting" reason="ShutdownRequested"
2025-06-09T19:45:59.698081Z DEBUG opentelemetry_sdk: name="BatchLogProcessor.ThreadStopped"
thread 'OpenTelemetry.Traces.BatchProcessor' has overflowed its stack
fatal runtime error: stack overflow
Note without debug logging, I see that other threads can also trigger the SO:
...
thread 'tokio-runtime-worker' has overflowed its stack
fatal runtime error: stack overflow
...
This makes me think that it has something to do with the the last caller to emit a log and/or a full channel, but that's just a guess.
The code that triggers it:
// let p: SdkLoggerProvider ...
...
p.force_flush().ok();
p.shutdown().ok();
Does not happen if i remove shutdown
.
Possibly related to #2897.
lean-apple
Metadata
Metadata
Assignees
Labels
No labels