Skip to content

fix(http_utils): disable httpx keepalive to spread load across uvicorn workers #29

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
rmfan wants to merge 1 commit into
base: prod
Choose a base branch
from
Open

fix(http_utils): disable httpx keepalive to spread load across uvicorn workers #29

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 11 additions & 1 deletion miles/utils/http_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -228,8 +228,18 @@ def init_http_client(args):

_client_concurrency = args.sglang_server_concurrency * args.rollout_num_gpus // args.rollout_num_gpus_per_engine
if _http_client is None:
# max_keepalive_connections=0: defeat connection reuse so each /run
# opens a fresh TCP. A pooled httpx.AsyncClient against a uvicorn
# multi-worker server pins all traffic to the few workers that
# originally accept()-won the pooled connections (uvicorn shares a
# single listen socket across --workers; no SO_REUSEPORT, no
# work-stealing). With keepalive off, every request runs its own
# accept() race and spreads across all workers.
_http_client = httpx.AsyncClient(
limits=httpx.Limits(max_connections=_client_concurrency),
limits=httpx.Limits(
max_connections=_client_concurrency,
max_keepalive_connections=0,
),
timeout=httpx.Timeout(None),
)

Expand Down
Loading