Hold off sending response to server until after first chunk to improve how errors are surfaced #108

Open
opened 2026-02-16 00:18:50 -05:00 by yindo · 2 comments
Owner

Originally created by @eyurtsev on GitHub (Jan 26, 2024).

Originally assigned to: @eyurtsev on GitHub.

Hold off sending a response from the server to the client for streaming end points until after the first chunk of the runnable was returned.

This allow sending back an actual http error to the client making errors more obvious

Originally created by @eyurtsev on GitHub (Jan 26, 2024). Originally assigned to: @eyurtsev on GitHub. Hold off sending a response from the server to the client for streaming end points until after the first chunk of the runnable was returned. This allow sending back an actual http error to the client making errors more obvious
yindo added the streaming label 2026-02-16 00:18:50 -05:00
Author
Owner

@eyurtsev commented on GitHub (Feb 6, 2024):

See this discussion: https://github.com/langchain-ai/langserve/discussions/433

@eyurtsev commented on GitHub (Feb 6, 2024): See this discussion: https://github.com/langchain-ai/langserve/discussions/433
Author
Owner

@ByPatrick commented on GitHub (Mar 13, 2024):

Hello, has there been any progress or plans regarding this issue? I am asking because I am encountering the same problem, where the client always receives a 500 error, but I would like to provide more details.

@ByPatrick commented on GitHub (Mar 13, 2024): Hello, has there been any progress or plans regarding this issue? I am asking because I am encountering the same problem, where the client always receives a 500 error, but I would like to provide more details.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langserve#108