Bug Fixes
- LLM Observability:
- Resolves an issue where a streamed Bedrock
InvokeModelWithResponseStreamrequest that ended in an error or was cancelled part-way produced a span with no output annotation. The text received before the failure is now tagged on the span. - Resolves an issue where a streamed Bedrock
ConverseStreamrequest that ended in an error or was cancelled part-way produced a span that was not marked as an error, making an interrupted request indistinguishable from a successful one. - This fix resolves an issue where cancelled or failed OpenAI streamed requests produced no LLM Observability span at all. These requests now produce an LLM span carrying the input messages, whatever output arrived before the stream ended, and estimated token metrics.
- Resolves an issue where a streamed Bedrock
- internal:
- Fixes incorrect telemetry being collected when hitting cardinality limits during stats computation.
- tracing:
- Fixes an issue where the Python garbage collector stops and memory grows without limit when
DD_RUNTIME_METRICS_ENABLEDis true.
- Fixes an issue where the Python garbage collector stops and memory grows without limit when
- Profiling:
- A crash that could occur in applications that use signal handlers to recover from crashes has been fixed.