apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-04-28 10:26:36 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix request closures during long-running streaming

Browse source
This commit is contained in:
adilhafeez 2026-04-18 18:10:23 -07:00
parent ffea891dba
commit c8c6b87d1e
7 changed files with 637 additions and 149 deletions
Download patch file Download diff file Expand all files Collapse all files

9
cli/planoai/utils.py
View file

@ -91,7 +91,12 @@ def convert_legacy_listeners(
"type": "model",
"port": 12000,
"address": "0.0.0.0",
"timeout": "30s",
# LLM streaming responses routinely exceed 30s (extended thinking,
# long tool reasoning, large completions). Match the 300s ceiling
# used by the direct upstream-provider routes so Envoy doesn't
# abort streams with UT mid-response. Users can override via their
# plano_config.yaml `listeners.timeout` field.
"timeout": "300s",
"model_providers": model_providers or [],
}
@ -100,7 +105,7 @@ def convert_legacy_listeners(
"type": "prompt",
"port": 10000,
"address": "0.0.0.0",
"timeout": "30s",
"timeout": "300s",
}
# Handle None case