Skip to content

fix(router): forward extra /sleep query params (level, mode) to upstream #969

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

Merged
ruizhang0101 merged 2 commits into from
Jun 19, 2026
+19 −6
Merged

fix(router): forward extra /sleep query params (level, mode) to upstream #969

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
74ebf63
fix(router): forward extra /sleep query params (level, mode) to upstream
HumphreySun98 Jun 4, 2026
f56a0c3
Merge branch 'main' into fix-sleep-forward-params
ruizhang0101 Jun 19, 2026
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
25 changes: 19 additions & 6 deletions src/vllm_router/services/request_service/request.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -517,9 +517,11 @@ async def route_general_request(
else:
endpoints = list(
filter(
lambda x: requested_model in x.model_names
and x.Id == request_endpoint
and not x.sleep,
lambda x: (
requested_model in x.model_names
and x.Id == request_endpoint
and not x.sleep
),
endpoints,
)
)
Expand Down Expand Up @@ -1064,21 +1066,32 @@ async def route_sleep_wakeup_request(

url = server_url + endpoint

# Forward any additional query parameters (e.g. /sleep `level` and `mode`,
# /wake_up `tags`) to the upstream engine. `id` is router-only and is
# consumed above to pick the target engine.
upstream_params = {k: v for k, v in request.query_params.items() if k != "id"}

@gemini-code-assist gemini-code-assist Bot Jun 4, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

Using a dictionary comprehension upstream_params = {k: v for k, v in request.query_params.items() if k != "id"} will drop duplicate query parameters (e.g., if multiple tags are provided like ?tags=a&tags=b, only the last one tags=b will be preserved). Since aiohttp supports a sequence of key-value tuples for the params argument, we can use a list comprehension with request.query_params.multi_items() to preserve all multi-valued query parameters.

Suggested change
upstream_params = {k: v for k, v in request.query_params.items() if k != "id"}
upstream_params = [(k, v) for k, v in request.query_params.multi_items() if k != "id"]


async with aiohttp.ClientSession() as client:
if endpoint == "/is_sleeping":
async with client.get(url, headers=headers) as response:
async with client.get(
url, headers=headers, params=upstream_params
) as response:
response.raise_for_status()
return await response.json()
else:
request_body = await request.body()
response_status = None
if request_body:
req_data = json.loads(request_body)
async with client.post(url, json=req_data, headers=headers) as response:
async with client.post(
url, json=req_data, headers=headers, params=upstream_params
) as response:
response.raise_for_status()
response_status = response.status
else:
async with client.post(url, headers=headers) as response:
async with client.post(
url, headers=headers, params=upstream_params
) as response:
response.raise_for_status()
response_status = response.status

Expand Down
Loading