Conversation
github-actions
bot
added
test:uipath-langchain
There was a problem hiding this comment.
Pull request overview
This PR adjusts the evaluation runtime’s logging verbosity so that eval lifecycle and diagnostic logs no longer appear at INFO level by default.
Changes:
- Downgraded multiple
logger.info(...)statements in the eval runtime tologger.debug(...). - Reduced default log noise for eval execution start/resume/suspend and trigger pass-through diagnostics.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| for i, trigger in enumerate(all_triggers, 1): | ||
| logger.info( | ||
| logger.debug( | ||
| f"EVAL RUNTIME: Pass-through trigger {i}: {trigger.model_dump(by_alias=True)}" | ||
| ) |
There was a problem hiding this comment.
These debug logs call trigger.model_dump(by_alias=True) inside an f-string. Even when debug logging is disabled, the f-string (and model_dump) will still be evaluated, which can add noticeable overhead if there are many triggers. Consider guarding this block with logger.isEnabledFor(logging.DEBUG) (or similar) so the model_dump only happens when debug logs will actually be emitted.
| logger.debug( | ||
| f"EVAL RUNTIME: Trigger {i}: {trigger.model_dump(by_alias=True)}" | ||
| ) |
There was a problem hiding this comment.
Same issue here: trigger.model_dump(by_alias=True) is computed unconditionally inside an f-string, even if debug logging is off. Please gate the model_dump behind a debug-level check to avoid unnecessary serialization work in normal runs.
| logger.debug( | |
| f"EVAL RUNTIME: Trigger {i}: {trigger.model_dump(by_alias=True)}" | |
| ) | |
| if logger.isEnabledFor(logging.DEBUG): | |
| trigger_dump = trigger.model_dump(by_alias=True) | |
| logger.debug( | |
| "EVAL RUNTIME: Trigger %s: %s", | |
| i, | |
| trigger_dump, | |
| ) |
| logger.debug( | ||
| f"DEBUG: Agent execution result status: {agent_execution_output.result.status}" | ||
| ) | ||
| logger.info( | ||
| logger.debug( | ||
| f"DEBUG: Agent execution result trigger: {agent_execution_output.result.trigger}" | ||
| ) |
There was a problem hiding this comment.
Now that these are emitted at debug level, the message prefix "DEBUG:" is redundant and makes log filtering/grepping noisier. Consider removing the "DEBUG:" prefix from these messages and keeping the rest of the context in the log text.
Description
This PR adjusts the evaluation runtime’s logging verbosity so that eval lifecycle and diagnostic logs no longer appear at INFO level by default.