mirror of
https://git.datalinker.icu/vllm-project/vllm.git
synced 2025-12-10 08:04:58 +08:00
decrement server_load on listen for disconnect (#18784)
Signed-off-by: Daniel Salib <danielsalib@meta.com>
This commit is contained in:
parent
435fa95444
commit
6e4cea1cc5
@ -26,6 +26,11 @@ async def listen_for_disconnect(request: Request) -> None:
|
||||
while True:
|
||||
message = await request.receive()
|
||||
if message["type"] == "http.disconnect":
|
||||
if request.app.state.enable_server_load_tracking:
|
||||
# on timeout/cancellation the BackgroundTask in load_aware_call
|
||||
# cannot decrement the server load metrics.
|
||||
# Must be decremented by with_cancellation instead.
|
||||
request.app.state.server_load_metrics -= 1
|
||||
break
|
||||
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user