starred/onyx
1
0 Fork 0
mirror of https://github.com/onyx-dot-app/onyx.git synced 2026-10-19 01:50:51 -05:00
Code Issues Packages Projects Releases 10 Wiki Activity

refactor(model-server): gate DISABLE_MODEL_SERVER in python entrypoint (#12754)

Browse Source 
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Jamison LahmanandClaude Opus 4.8 authored and GitHub committed 2026-07-07 23:48:13 +00:00
1 parent df4ab53a8b
commit d557bd3702
11 files changed
+42 -101
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

backend/Dockerfile.model_server
+1 -1
View File
@@ -82,4 +82,4 @@ ENV PYTHONPATH=/app
ARG ONYX_VERSION=0.0.0-dev
ENV ONYX_VERSION=${ONYX_VERSION}
CMD ["uvicorn", "model_server.main:app", "--host", "0.0.0.0", "--port", "9000"]
CMD ["python", "-m", "model_server"]
backend/model_server/__main__.py
+29
View File
@@ -0,0 +1,29 @@
"""Process entry point for the model server (`python -m model_server`).
The `DISABLE_MODEL_SERVER` gate runs here, ahead of `model_server.main`'s heavy ML
imports, so a disabled container exits without loading torch / the model stack.
"""
import sys
from onyx.utils.logger import setup_logger
from shared_configs.configs import DISABLE_MODEL_SERVER
logger = setup_logger()
def main() -> None:
if DISABLE_MODEL_SERVER:
# The deployment points inference/indexing at an external model server, so this
# container has nothing to run. Exit cleanly instead of starting uvicorn.
logger.notice("DISABLE_MODEL_SERVER is set; skipping model server startup.")
sys.exit(0)
# Imported lazily so the disabled path above stays free of the heavy ML imports.
from model_server.main import run_server
run_server()
if __name__ == "__main__":
main()
backend/model_server/main.py
+12 -4
View File
@@ -24,7 +24,6 @@ from onyx.utils.middleware import add_onyx_request_id_middleware
from onyx.utils.middleware import add_onyx_tenant_id_middleware
from shared_configs.configs import INDEXING_ONLY
from shared_configs.configs import MIN_THREADS_ML_MODELS
from shared_configs.configs import MODEL_SERVER_ALLOWED_HOST
from shared_configs.configs import MODEL_SERVER_PORT
from shared_configs.configs import SENTRY_DSN
from shared_configs.configs import SENTRY_TRACES_SAMPLE_RATE
@@ -141,11 +140,20 @@ def get_model_app() -> FastAPI:
app = get_model_app()
if __name__ == "__main__":
def run_server() -> None:
# Bind all interfaces (loopback included) so the container healthcheck's
# localhost probe reaches the server. Mirrors the old shell command's
# `--host 0.0.0.0`; MODEL_SERVER_HOST is a client-side address and must not
# drive the bind host.
host = "0.0.0.0" # noqa: S104
logger.notice(
"Starting Onyx Model Server on http://%s:%s/",
MODEL_SERVER_ALLOWED_HOST,
host,
str(MODEL_SERVER_PORT),
)
logger.notice("Model Server Version: %s", __version__)
uvicorn.run(app, host=MODEL_SERVER_ALLOWED_HOST, port=MODEL_SERVER_PORT)
uvicorn.run(app, host=host, port=MODEL_SERVER_PORT)
if __name__ == "__main__":
run_server()
deployment/aws_ecs_fargate/cloudformation/services/onyx_model_server_indexing_service_template.yaml
-6
View File
@@ -120,12 +120,6 @@ Resources:
Image: onyxdotapp/onyx-model-server:latest
Cpu: 0
Essential: true
Command:
- "/bin/sh"
- "-c"
- >
if [ "${DISABLE_MODEL_SERVER}" = "True" ] || [ "${DISABLE_MODEL_SERVER}" = "true" ]; then echo 'Skipping service...';
exit 0; else exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000; fi
PortMappings:
- Name: model-server
ContainerPort: 9000
deployment/aws_ecs_fargate/cloudformation/services/onyx_model_server_inference_service_template.yaml
-6
View File
@@ -120,12 +120,6 @@ Resources:
Image: onyxdotapp/onyx-model-server:latest
Cpu: 0
Essential: true
Command:
- "/bin/sh"
- "-c"
- >
if [ "${DISABLE_MODEL_SERVER}" = "True" ] || [ "${DISABLE_MODEL_SERVER}" = "true" ]; then echo 'Skipping service...';
exit 0; else exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000; fi
PortMappings:
- Name: model-server
ContainerPort: 9000
deployment/docker_compose/docker-compose.multitenant-dev.yml
-14
View File
@@ -350,13 +350,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
@@ -379,13 +372,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- INDEX_BATCH_SIZE=${INDEX_BATCH_SIZE:-}
deployment/docker_compose/docker-compose.prod-cloud.yml
-14
View File
@@ -164,13 +164,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
@@ -190,13 +183,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
deployment/docker_compose/docker-compose.prod-no-letsencrypt.yml
-14
View File
@@ -165,13 +165,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
@@ -193,13 +186,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
deployment/docker_compose/docker-compose.prod.yml
-14
View File
@@ -201,13 +201,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: unless-stopped
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
@@ -229,13 +222,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: unless-stopped
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
deployment/docker_compose/docker-compose.search-testing.yml
-14
View File
@@ -113,13 +113,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
@@ -137,13 +130,6 @@ services:
build:
context: ../../backend
dockerfile: Dockerfile.model_server
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
restart: on-failure
environment:
- MIN_THREADS_ML_MODELS=${MIN_THREADS_ML_MODELS:-}
deployment/docker_compose/docker-compose.yml
-14
View File
@@ -274,13 +274,6 @@ services:
# - driver: nvidia
# count: all
# capabilities: [gpu]
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
env_file:
- path: .env
required: false
@@ -324,13 +317,6 @@ services:
# - driver: nvidia
# count: all
# capabilities: [gpu]
command: >
/bin/sh -c "if [ \"${DISABLE_MODEL_SERVER:-}\" = \"True\" ] || [ \"${DISABLE_MODEL_SERVER:-}\" = \"true\" ]; then
echo 'Skipping service...';
exit 0;
else
exec uvicorn model_server.main:app --host 0.0.0.0 --port 9000;
fi"
env_file:
- path: .env
required: false