Make port 30101 work everywhere (#134)

This PR changes edge-endpoints default port to 30101 so that the view
from outside k3s is the same as the view from direct-to-docker. I've
also updated comments, readmes, etc.

.github/workflows/pipeline.yaml

@@ -15,7 +15,7 @@ env:
   # This is the token associated with "prod-biggies" (with shared credentials on 1password)
   GROUNDLIGHT_API_TOKEN: ${{ secrets.GROUNDLIGHT_API_TOKEN }}
   # This is the NGINX proxy endpoint
-  GROUNDLIGHT_ENDPOINT: http://localhost:6717
+  GROUNDLIGHT_ENDPOINT: http://localhost:30101
 
 jobs:
   test-general-edge-endpoint:
@@ -79,13 +79,13 @@ jobs:
           container_id=$(docker run \
             -e LOG_LEVEL=DEBUG \
             -e EDGE_CONFIG \
-            -d -p 6717:6717 \
+            -d -p 30101:30101 \
             groundlight-edge)
           echo "::set-output name=container_id::$container_id"
 
       - name: Run Unit Tests (that do require docker)
         run: |
-          GROUNDLIGHT_ENDPOINT=http://localhost:6717
+          GROUNDLIGHT_ENDPOINT=http://localhost:30101
           source test/setup_plain_test_env.sh
           poetry run pytest -k "_live"
 
@@ -132,7 +132,7 @@ jobs:
           container_id=$(docker run \
             -e LOG_LEVEL=DEBUG \
             -e EDGE_CONFIG \
-            -d -p 6717:6717 \
+            -d -p 30101:30101 \
             groundlight-edge)
           echo "::set-output name=container_id::$container_id"
 

Dockerfile

@@ -1,5 +1,6 @@
 # Build args
-ARG NGINX_PORT=6717
+ARG NGINX_PORT=30101
+ARG NGINX_PORT_OLD=6717
 ARG UVICORN_PORT=6718
 ARG APP_ROOT="/groundlight-edge"
 ARG POETRY_HOME="/opt/poetry"
@@ -23,12 +24,12 @@ ARG POETRY_VERSION
 # since they are required by OpenCV
 RUN apt-get update \
     && apt-get install -y --no-install-recommends \
-       bash \
-       curl \
-       nginx \
-       libglib2.0-0 \
-       libgl1-mesa-glx \
-       sqlite3 \
+    bash \
+    curl \
+    nginx \
+    libglib2.0-0 \
+    libgl1-mesa-glx \
+    sqlite3 \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/* \
     && curl -sSL https://install.python-poetry.org | python -
@@ -95,4 +96,4 @@ RUN ln -sf /dev/stdout /var/log/nginx/access.log && \
 CMD nginx && poetry run uvicorn --workers 8 --host 0.0.0.0 --port ${APP_PORT} --proxy-headers app.main:app
 
 # Document the exposed port, which is configured in nginx.conf
-EXPOSE ${NGINX_PORT}
+EXPOSE ${NGINX_PORT} ${NGINX_PORT_OLD}

README.md

@@ -26,7 +26,7 @@ export GROUNDLIGHT_ENDPOINT=http://localhost:30101
 To find the correct port, run `kubectl get services` and you should see an entry like this:
 ```
 NAME                                                        TYPE       CLUSTER-IP      EXTERNAL-IP   PORT(S)                         AGE
-service/edge-endpoint-service                               NodePort   10.43.141.253   <none>        6717:30101/TCP                  23m
+service/edge-endpoint-service                               NodePort   10.43.141.253   <none>        30101:30101/TCP                 23m
 ```
 
 The port is the second number listed under ports for the `edge-endpoint-service` (in this case, 30101).

configs/nginx.conf

@@ -12,7 +12,8 @@ http {
     access_log /var/log/nginx/access.log custom_combined;
 
     server {
-        listen 6717;
+        listen 30101;
+        listen 6717; # This is the old port, which we should take out once we are sure that folks are all using the new port.
         client_max_body_size 10M;   # a bit more than enough for a large json of a 14MP image
 
         location / {

deploy/README.md

@@ -5,13 +5,6 @@ The edge endpoint runs under kubernetes, typically on a single-node cluster, whi
 
 The instructions below are fairly opinionated, optimized for single-node cluster setup, using k3s, on an Ubuntu/Debian-based system.  If you want to set it up with a different flavor of kubernetes, that should work, but you'll have to figure out how to do that yourself.
 
-## Dependencies
-Edge Endpoint requires `curl` and `jq` which are likely already on your system, but if you have a minimal Linux distribution, they might not be. To make sure you have them, run:
-
-```shell
-sudo apt update && sudo apt install -y jq curl
-```
-
 ## Setting up Single-Node Kubernetes with k3s
 
 If you don't have [k3s](https://docs.k3s.io/) installed, there are two scripts which can install it depending on whether you have a CUDA GPU or not.  If you don't set up a GPU, the models will run more slowly on CPU.
@@ -26,6 +19,8 @@ If you don't have [k3s](https://docs.k3s.io/) installed, there are two scripts w
 ./deploy/bin/install-k3s-nvidia.sh
 ```
 
+(Note: these scripts depend on the Linux utilities `curl` and `jq`. If these aren't on your system, the scripts will install them for you.)
+
 You might want to customize the [edge config file](../configs/edge-config.yaml) to include the detector ID's you want to run.  Adding detector ID's to the config file will cause inference pods to be initialized automatically for each detector. Even if they aren't configured in the config file,
 edge inference will be set up for each detector ID for which the Groundlight service receives requests (note that it
 takes some time for each inference pod to become available for the first time).

deploy/k3s/edge_deployment/edge_deployment.yaml

@@ -15,7 +15,7 @@ spec:
   ports:
   - protocol: TCP
     # Service port for NGINX
-    port: 6717
+    port: 30101
     nodePort: 30101
   type: NodePort
 ---
@@ -54,7 +54,7 @@ spec:
         image: *edgeEndpointImage
         imagePullPolicy: Always
         ports:
-        - containerPort: 6717
+        - containerPort: 30101
         env:
         - name: LOG_LEVEL
           value: "INFO"
@@ -75,14 +75,14 @@ spec:
         startupProbe:
           httpGet:
             path: /health/live  # Checks if the server is up
-            port: 6717
+            port: 30101
           initialDelaySeconds: 10
           periodSeconds: 10
           failureThreshold: 6  # Wait for up to 1 min
         readinessProbe:
           httpGet:
             path: /health/ready  # Checks if the server is ready to serve requests
-            port: 6717
+            port: 30101
           initialDelaySeconds: 10
           periodSeconds: 10
           failureThreshold: 6  # Wait for up to 1 min

test/api/test_image_queries_live.py

@@ -11,7 +11,7 @@
 
 # Tests in this file require a live edge-endpoint server and GL Api token in order to run.
 # Not ideal for unit-testing.
-TEST_ENDPOINT = "http://localhost:6717"
+TEST_ENDPOINT = "http://localhost:30101"
 MAX_WAIT_TIME_S = 60
 
 # Detector ID associated with the detector with parameters

test/setup_plain_test_env.sh

@@ -13,7 +13,7 @@
 # > docker run --name groundlight-edge \
 #      -e LOG_LEVEL=DEBUG \
 #      -e EDGE_CONFIG \
-#      --rm -it -p 6717:6717 edge-endpoint
+#      --rm -it -p 30101:30101 edge-endpoint
 
 # Then in another terminal, run the tests:
 # > make test-with-docker