Improve LiteLLM key rotation and monitoring #3
Description
Problem
During the Sept 30 Agentic Workshop, the LiteLLM authentication key briefly expired (few seconds), causing temporary service disruption.
Context
Currently using open-source LiteLLM which has limited key management flexibility. Enterprise version not justified for current needs.
Requirements
- Implement monitoring/alerting for LiteLLM key expiration (warning at 7 days, 3 days, 1 day before expiry)
- Document key rotation procedure
- Add key expiration check to pre-workshop validation checklist
- Investigate if open-source version supports any form of key rotation without downtime
- Document decision criteria for when Enterprise LiteLLM would be justified
Success Criteria
- Proactive alerting prevents key expiration during workshops
- Key rotation can be performed without service disruption
- Pre-workshop checklist catches expiring keys before events
Related
Sept 30 Workshop Postmortem