Prevent active chat streams from being cancelled after 60 seconds #68
Reference in New Issue
Block a user
Delete Branch "fix/stream-stall-timeout"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Quart applies a 60-second response deadline to the chat SSE endpoint, cancelling active generation even when the model continues producing output. Disable the response deadline for this endpoint while preserving the existing silent-model stall timeout.
Add a regression test that streams beyond a shortened Quart response deadline and still completes successfully.
Validation: all 6 tests in tests/unit/test_stream_stall.py pass; Ruff and formatting hooks pass.