Working with Analysis Results and Common Issues

The Log Analysis Agent not only identifies operational failures, but also supports ongoing troubleshooting workflows through follow-up analysis, KEDB integration, and structured remediation guidance.

This article explains how to work with analysis results, save new findings, and resolve common operational issues encountered while using the agent.

Example Requests

Analysing Pasted Logs

Here's an error from yesterday's run, can you analyse it?

[paste full stack trace]

Analysing Local Logs

Analyse the log at /var/log/syntasa/job-78321.log

Analysing Cloud Storage Logs

Diagnose errors in gs://syntasa-prod-logs/ingest/2026-05-08/run.log
Look at this log for me: s3://customer-logs/etl-failed-2026-05-07.log

Follow-Up Questions

After the initial analysis, users can continue asking operational follow-up questions in the same session without re-uploading logs.

Examples include:

  • “What does error #2 mean in plain English?”
  • “Which of these errors should I fix first?”
  • “What’s the permanent fix for the OOM error?”

Saving Unknown Errors

When the agent identifies previously unknown errors, it may prompt users to save them into the KEDB.

Supported responses include:

Save
yes, save them

 

This helps improve future automated troubleshooting for similar incidents.

Providing Operational Feedback

Users can provide feedback on KEDB recommendations directly within the chat session.

Examples:

The fix for KE-2104 worked, thanks
The workaround for KE-1789 didn't work

 

This feedback helps improve the quality and accuracy of future recommendations.

Common Issues and Resolutions

Off-Topic Requests

If the agent responds with:

“I'm sorry, but I can only assist with Syntasa log analysis…”

the request was likely outside the supported troubleshooting scope. Users should either:

  • Provide a log or log path
  • Switch to the Syntasa Help Agent for platform guidance

Permission Errors for GCS or S3

Errors such as:

  • “Permission denied”
  • “I couldn't fetch the file”

typically indicate that the platform service account lacks access to the bucket.

Recommended solutions:

  • Paste the logs directly into chat
  • Grant the syntasa-ai service account read access to the bucket

Agent Re-Asks for the Log

This usually occurs when follow-up questions are asked in a new chat session.

The agent only retains prior analyses within the same session.

Unknown Error Detection

If the report shows:

kedb_match: false

for a known issue, it may indicate that the wording of the error differs from existing KEDB entries.

Possible solutions include:

  • Re-running the search with the exact root cause phrasing
  • Saving the new variant as a separate KEDB entry

Empty or Malformed Analysis Output

This can happen when extremely large or truncated logs are provided.

Recommended solution:

  • Re-send a smaller, focused log segment containing the relevant error blocks

Conclusion

The Log Analysis Agent supports a complete troubleshooting workflow — from initial error detection to structured remediation guidance and KEDB knowledge expansion. By understanding how to interact with the agent effectively and resolve common operational issues, teams can significantly improve troubleshooting efficiency and operational reliability.