mirror of
https://github.com/Significant-Gravitas/AutoGPT.git
synced 2026-04-08 03:00:28 -04:00
abe3707f17
Replace in-memory _running_operations dict with Redis-based tracking
to support Kubernetes deployments with multiple pods. The in-memory
dict only tracked operations per-process, allowing duplicate operations
to start if requests hit different pods.
Changes:
- Add _mark_operation_started() using Redis SETNX for atomic idempotency
- Add _mark_operation_completed() to clear Redis key on completion
- Use Redis key "chat:running_operation:{tool_call_id}" with 10min TTL
- Update _execute_long_running_tool() to clear key in finally block