Skip to content

Propagate extra server args to the gunicorn command #291

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
merged 3 commits into from
Sep 26, 2023
Merged

Propagate extra server args to the gunicorn command #291

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8905cc7
Propagate extra server args to the gunicorn command
squeakymouse Sep 26, 2023
4497180
default graceful-timeout to 600
squeakymouse Sep 26, 2023
49a56d4
undo changes for gateway
squeakymouse Sep 26, 2023
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
3 changes: 2 additions & 1 deletion model-engine/model_engine_server/inference/forwarding/echo_server.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@ def entrypoint():
parser.add_argument("--host", type=str, default="[::]")
parser.add_argument("--port", type=int, default=5009)

args = parser.parse_args()
args, extra_args = parser.parse_known_args()

command = [
"gunicorn",
Expand All @@ -48,6 +48,7 @@ def entrypoint():
"--workers",
str(args.num_workers),
"model_engine_server.inference.forwarding.echo_server:app",
*extra_args,
]
subprocess.run(command)

Expand Down
6 changes: 5 additions & 1 deletion model-engine/model_engine_server/inference/forwarding/http_forwarder.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,8 +138,9 @@ def entrypoint():
parser.add_argument("--host", type=str, default="[::]")
parser.add_argument("--port", type=int, default=5000)
parser.add_argument("--set", type=str, action="append")
parser.add_argument("--graceful-timeout", type=int, default=600)

args = parser.parse_args()
args, extra_args = parser.parse_known_args()

values = [f"CONFIG_FILE={args.config}"]
if args.set is not None:
Expand All @@ -160,8 +161,11 @@ def entrypoint():
"uvicorn.workers.UvicornWorker",
"--workers",
str(args.num_workers),
"--graceful-timeout",
str(args.graceful_timeout),
*envs,
"model_engine_server.inference.forwarding.http_forwarder:app",
*extra_args,
Copy link
Contributor

@seanshi-scale seanshi-scale Sep 26, 2023

Choose a reason for hiding this comment

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

oh wait hmm I think it might be good to just hardcode in the termination grace period seconds as a default here idk

Copy link
Contributor

@seanshi-scale seanshi-scale Sep 26, 2023

Choose a reason for hiding this comment

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

extra_args feels like it should be a dict to me (e.g. --key1 val1 --key2 val2 from {key1: val1, key2: val2}), is there such thing as extra_kwargs?

this way we can set a default of graceful-timeout and override it if necessary

Copy link
Contributor Author

@squeakymouse squeakymouse Sep 26, 2023

Choose a reason for hiding this comment

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

Ah, so default it to 600 seconds? Will do! :)

Copy link
Contributor Author

@squeakymouse squeakymouse Sep 26, 2023

Choose a reason for hiding this comment

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

Oh wait didn't see the second comment when I commented ^that (GitHub is bad at refreshing lol 😅 ). From local testing, parser.parse_known_args() will turn --key1 val1 --key2 val2 into extra_args = ["--key1", "val1", "--key2", "val2"], which seems fine to directly pass? :P

Copy link
Contributor

@seanshi-scale seanshi-scale Sep 26, 2023

Choose a reason for hiding this comment

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

yeah it's fine to pass; I was hoping for a dict but if we don't get that from argparse then we can just directly pass it

]
subprocess.run(command)

Expand Down
8 changes: 8 additions & 0 deletions model-engine/model_engine_server/inference/sync_inference/start_fastapi_server.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,4 @@
import argparse
import os
import subprocess

Expand All @@ -8,6 +9,10 @@


def start_server():
parser = argparse.ArgumentParser()
parser.add_argument("--graceful-timeout", type=int, default=600)
args, extra_args = parser.parse_known_args()

# TODO: HTTPS
command = [
"gunicorn",
Expand All @@ -21,7 +26,10 @@ def start_server():
"uvicorn.workers.UvicornWorker",
"--workers",
str(NUM_PROCESSES),
"--graceful-timeout",
str(args.graceful_timeout),
"model_engine_server.inference.sync_inference.fastapi_server:app",
*extra_args,
]
unset_sensitive_envvars()
subprocess.run(command)
Expand Down