fix: redact sensitive data from evaluation SDK log messages#45176
Open
slister1001 wants to merge 2 commits intoAzure:mainfrom
Open
fix: redact sensitive data from evaluation SDK log messages#45176slister1001 wants to merge 2 commits intoAzure:mainfrom
slister1001 wants to merge 2 commits intoAzure:mainfrom
Conversation
github-actions
bot
added
the
Evaluation
Contributor
There was a problem hiding this comment.
Pull request overview
This PR addresses a critical security issue by redacting sensitive user-provided data from log messages in the azure-ai-evaluation SDK that flow into Geneva telemetry. The changes prevent database connection strings and other credentials embedded in user payloads from being exposed in logs and flagged by CredScan tools.
Changes:
- Removed user-provided content (queries, responses, tool definitions) from log message strings
- Downgraded noisy agent-response fallback warnings from
logger.warningtologger.debug - Sanitized upload error messages to emit only exception type names instead of full exception text
- Added proper exception chaining with
from efor maintaining traceback while keeping logs clean
| return response | ||
| return "\n".join(agent_response) | ||
| except: | ||
| # If the agent response cannot be parsed for whatever reason (e.g. the converter format changed), the original response is returned |
There was a problem hiding this comment.
The bare except: clause should specify Exception to avoid catching system-exiting exceptions like KeyboardInterrupt and SystemExit. This is inconsistent with the exception handling patterns used elsewhere in the file (see lines 700 and 786 which use except Exception as e:). Consider changing this to except Exception: for consistency and better practice.
sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_common/utils.py
Outdated
Show resolved
Hide resolved
slister1001
force-pushed
the
fix/credscan-redact-logs
branch
from
b6281bd to
2092f7c
Compare
Remove user-provided content (queries, responses, tool definitions, exception messages) from log strings that flow into Geneva telemetry. This prevents database connection strings and other credentials embedded in user payloads from being flagged by CredScan. Changes: - Remove f-string interpolation of query/response/tool_definitions in warning and debug log messages - Downgrade noisy agent-response warnings to debug level - Sanitize upload error messages to emit only exception type name - Chain original exception with 'from e' Resolves ICM 738457593
slister1001
force-pushed
the
fix/credscan-redact-logs
branch
from
2092f7c to
b5b0674
Compare
PROXY_URL in devtools_testutils.config is now a function, not a constant. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Description
Redacts user-provided content (queries, responses, tool definitions, exception messages) from log strings in
azure-ai-evaluationthat flow into Geneva telemetry (ViennaExp/Log). This prevents database connection strings and other credentials embedded in user payloads from being flagged by CredScan.Resolves ICM 738457593
Changes
logger.debugtype(e).__name__instead of full exception textfrom efor proper tracebackAll SDK Contribution checklist:
General Guidelines and Best Practices
Testing Guidelines