Include Production WSGI (Gunicorn) to replace Flask Default Server

Makefile

@@ -5,6 +5,8 @@ IMAGE_VERSION := 0.7
 IMAGE ?= $(IMAGE_REGISTRY)/$(IMAGE_NAME):v$(IMAGE_VERSION)
 LATEST_TAG_IMAGE := $(IMAGE_REGISTRY)/$(IMAGE_NAME):latest
 TEST_IMAGE := $(IMAGE)-test
+MODEL_SERVER_TEST_IMAGE := $(TEST_IMAGE)-model-server
+OFFLINE_TRAINER_TEST_IMAGE := $(TEST_IMAGE)-offline-trainer
 
 CTR_CMD = docker
 
@@ -20,6 +22,12 @@ build-test-nobase:
 build-test:
  $(CTR_CMD) build -t $(TEST_IMAGE) -f $(DOCKERFILES_PATH)/Dockerfile.test .
 
+build-test-model-server:
+ $(CTR_CMD) build -t $(MODEL_SERVER_TEST_IMAGE) -f $(DOCKERFILES_PATH)/Dockerfile.test-model-server .
+
+build-test-offline-trainer:
+ $(CTR_CMD) build -t $(OFFLINE_TRAINER_TEST_IMAGE) -f $(DOCKERFILES_PATH)/Dockerfile.test-offline-trainer .
+
 push:
  $(CTR_CMD) push $(IMAGE)
 
@@ -51,28 +59,75 @@ run-model-server:
  $(CTR_CMD) run -d --platform linux/amd64 -e "MODEL_TOPURL=http://localhost:8110" -v ${MODEL_PATH}:/mnt/models -p 8100:8100 --name model-server $(TEST_IMAGE) /bin/bash -c "python3.8 tests/http_server.py & sleep 10 && python3.8 src/server/model_server.py"
  while ! docker logs model-server | grep -q Serving; do echo "waiting for model-server to serve"; sleep 5; done
 
+run-model-server-prod:
+ $(CTR_CMD) run -d \
+ --platform linux/amd64 \
+ -e "MODEL_TOPURL=http://localhost:8110" \
+ -v ${MODEL_PATH}:/mnt/models \
+ -p 8105:8105 \
+ --name model-server-prod \
+ $(MODEL_SERVER_TEST_IMAGE)
+
 run-estimator-client:
  $(CTR_CMD) exec model-server /bin/bash -c "python3.8 -u ./tests/estimator_model_request_test.py"
 
 clean-model-server:
  @$(CTR_CMD) stop model-server
  @$(CTR_CMD) rm model-server
 
-test-model-server: run-model-server run-estimator-client clean-model-server
+clean-model-server-prod:
+ @$(CTR_CMD) stop model-server-prod
+ @$(CTR_CMD) rm model-server-prod
+
+test-model-server: 
+ @if [ "$(ENV)" = "prod" ]; then \
+ run-model-server-prod run-estimator-client clean-model-server-prod
+ else \
+ run-model-server run-estimator-client clean-model-server
+ fi
 
 # test offline trainer
 run-offline-trainer:
- $(CTR_CMD) run -d --platform linux/amd64 -p 8102:8102 --name offline-trainer $(TEST_IMAGE) python3.8 src/train/offline_trainer.py
+ $(CTR_CMD) run -d --platform linux/amd64 \
+ -p 8102:8102 \ 
+ --name offline-trainer \ 
+ $(TEST_IMAGE) \ 
+ python3.8 src/train/offline_trainer.py
  sleep 5
 
 run-offline-trainer-client:
- $(CTR_CMD) exec offline-trainer /bin/bash -c "python3.8 -u ./tests/offline_trainer_test.py"
+ $(CTR_CMD) exec offline-trainer \
+ /bin/bash -c \
+ "python3.8 -u ./tests/offline_trainer_test.py"
+
+
+run-offline-trainer-client-prod:
+ $(CTR_CMD) exec offline-trainer-prod \
+ /bin/bash -c \
+ "python3.8 -u ./tests/offline_trainer_test.py"
+
+run-offline-trainer-prod:
+ $(CTR_CMD) run -d \
+ --platform linux/amd64 \
+ -p 8107:8107 \
+ --name offline-trainer-prod \
+ $(OFFLINE_TRAINER_TEST_IMAGE)
+ sleep 5
 
 clean-offline-trainer:
  @$(CTR_CMD) stop offline-trainer
  @$(CTR_CMD) rm offline-trainer
 
-test-offline-trainer: run-offline-trainer run-offline-trainer-client clean-offline-trainer
+clean-offline-trainer-prod:
+ @$(CTR_CMD) stop offline-trainer-prod
+ @$(CTR_CMD) rm offline-trainer-prod
+
+test-offline-trainer: 
+ @if [ "$(ENV)" = "prod" ]; then \
+ run-offline-trainer-prod run-offline-trainer-client-prod clean-offline-trainer-prod
+ else \
+ run-offline-trainer run-offline-trainer-client clean-offline-trainer
+ fi
 
 test: build-test test-pipeline test-estimator test-model-server test-offline-trainer
 

config/gunicorn_config.py

@@ -0,0 +1,9 @@
+import os
+
+
+workers = int(os.environ.get('GUNICORN_PROCESSES', '2'))
+threads = int(os.environ.get('GUNICORN_THREADS', '4'))
+port = os.environ.get('GUNICORN_PORT', '8100')
+bind = os.environ.get('GUNICORN_BIND', '0.0.0.0:' + port)
+forwarded_allow_ips = '*'
+secure_scheme_headers = { 'X-Forwarded-Proto': 'https' }

dockerfiles/Dockerfile

@@ -12,10 +12,13 @@ COPY src/util src/util
 COPY cmd cmd
 
 # port for Model Server
-EXPOSE 8100
+ENV GUNICORN_PORT=8105
+
+EXPOSE ${GUNICORN_PORT}
 # port for Online Trainer (TODO: reserved for event-based online training)
-EXPOSE 8101
+#EXPOSE 8101
 # port for Offline Trainer
-EXPOSE 8102
+#EXPOSE 8102
+
 
-ENTRYPOINT ["python3.8", "cmd/main.py"]
+CMD ["gunicorn", "--config", "config/gunicorn_config.py", "src.server.model_server:app"]

dockerfiles/Dockerfile.test

@@ -24,4 +24,4 @@ EXPOSE 8101
 # port for Offline Trainer
 EXPOSE 8102
 
-CMD [ "python3.8", "-u", "src/server/model_server.py" ]
+CMD [ "python3.8", "-u", "src/server/model_server.py" ]

dockerfiles/Dockerfile.test-model-server

@@ -0,0 +1,26 @@
+FROM --platform=linux/amd64 quay.io/sustainable_computing_io/kepler_model_server_base:v0.7
+
+WORKDIR /usr/local
+
+RUN mkdir -p /usr/local/src
+RUN mkdir -p /usr/local/resource
+
+COPY dockerfiles/requirements.txt dockerfiles/requirements.txt
+RUN pip install --no-cache-dir -r dockerfiles/requirements.txt
+
+COPY src/estimate src/estimate
+COPY src/server src/server
+COPY src/train src/train
+COPY src/util src/util
+COPY cmd cmd
+COPY config config
+
+RUN mkdir -p tests/data
+COPY tests/data/prom_output tests/data/prom_output 
+COPY tests/*.py tests/
+
+ENV GUNICORN_PORT=8105
+
+EXPOSE ${GUNICORN_PORT}
+
+CMD ["gunicorn", "--config", "config/gunicorn_config.py", "src.server.model_server:app"]

dockerfiles/Dockerfile.test-offline-trainer

@@ -0,0 +1,26 @@
+FROM --platform=linux/amd64 quay.io/sustainable_computing_io/kepler_model_server_base:v0.7
+
+WORKDIR /usr/local
+
+RUN mkdir -p /usr/local/src
+RUN mkdir -p /usr/local/resource
+
+COPY dockerfiles/requirements.txt dockerfiles/requirements.txt
+RUN pip install --no-cache-dir -r dockerfiles/requirements.txt
+
+COPY src/estimate src/estimate
+COPY src/server src/server
+COPY src/train src/train
+COPY src/util src/util
+COPY cmd cmd
+COPY config config
+
+RUN mkdir -p tests/data
+COPY tests/data/prom_output tests/data/prom_output 
+COPY tests/*.py tests/
+
+ENV GUNICORN_PORT=8107
+
+EXPOSE ${GUNICORN_PORT}
+
+CMD ["gunicorn", "--config", "config/gunicorn_config.py", "src.train.offline_trainer:app"]

dockerfiles/requirements.txt

@@ -1,4 +1,5 @@
 flask==2.1.2
+gunicorn==22.0.0
 Werkzeug==2.2.2
 protobuf==3.19.4
 pandas==1.4.4