fix: ingestion docker image (#2027)

The environment was not set correctly, so it could not fire kafka events. It (mce-cli) always worked when running outside of docker.

I also added a dev ingestion docker image / script which may be much faster if you've already built locally.

Tested:
1. Cleaned docker volumes and started datahub. Verified it is empty.
2. Built with gradle.
3. Ran ./docker/ingestion/ingestion-dev.sh. Verified data shows in DataHub.
4. Ran step 1 again.
5. Ran ./docker/ingestion/ingestion.sh. Verified data shows in DataHub.

docker/ingestion/Dockerfile

@@ -1,8 +1,19 @@
-FROM openjdk:8 as builder
+# Defining environment
+ARG APP_ENV=prod
+
+FROM openjdk:8-jre-alpine as base
+
+FROM openjdk:8 as prod-build
 COPY . datahub-src
 RUN cd datahub-src && ./gradlew :metadata-ingestion-examples:mce-cli:build
 
-FROM openjdk:8-jre-alpine
-COPY --from=builder datahub-src/metadata-ingestion-examples/mce-cli/build/libs/mce-cli.jar ./
-COPY --from=builder datahub-src/metadata-ingestion-examples/mce-cli/example-bootstrap.json ./
-CMD java -jar mce-cli.jar -m produce example-bootstrap.json
+FROM base as prod-install
+COPY --from=prod-build datahub-src/metadata-ingestion-examples/mce-cli/build/libs/mce-cli.jar /datahub/ingestion/bin/mce-cli.jar
+COPY --from=prod-build datahub-src/metadata-ingestion-examples/mce-cli/example-bootstrap.json /datahub/ingestion/example-bootstrap.json
+
+FROM base as dev-install
+# Dummy stage for development. Assumes code is built on your machine and mounted to this image.
+# See this excellent thread https://github.com/docker/cli/issues/1134
+
+FROM ${APP_ENV}-install as final
+CMD java -jar /datahub/ingestion/bin/mce-cli.jar -m produce /datahub/ingestion/example-bootstrap.json

docker/ingestion/docker-compose.dev.yml

@@ -0,0 +1,18 @@
+---
+version: '3.5'
+services:
+  ingestion:
+    image: datahub-ingestion:debug
+    env_file: env/docker.env
+    build:
+      context: .
+      dockerfile: Dockerfile
+      args:
+        APP_ENV: dev
+    volumes:
+      - ../../metadata-ingestion-examples/mce-cli/build/libs/:/datahub/ingestion/bin
+      - ../../metadata-ingestion-examples/mce-cli/example-bootstrap.json:/datahub/ingestion/example-bootstrap.json
+
+networks:
+  default:
+    name: datahub_network

docker/ingestion/docker-compose.yml

@@ -3,6 +3,7 @@ version: '3.5'
 services:
   ingestion:
     image: datahub-ingestion
+    env_file: env/docker.env
     build:
       context: ../../
       dockerfile: docker/ingestion/Dockerfile

docker/ingestion/env/docker.env

@@ -0,0 +1,2 @@
+KAFKA_BOOTSTRAP_SERVER=broker:29092
+KAFKA_SCHEMAREGISTRY_URL=http://schema-registry:8081

docker/ingestion/ingestion-dev.sh

@@ -0,0 +1,6 @@
+#!/bin/bash
+
+# Runs the ingestion image using your locally built mce-cli. Gradle build must have been run before this script.
+
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
+cd $DIR && COMPOSE_DOCKER_CLI_BUILD=1 DOCKER_BUILDKIT=1 docker-compose -f docker-compose.dev.yml -p datahub up

metadata-ingestion-examples/mce-cli/README.md

@@ -48,4 +48,11 @@ java -jar  metadata-ingestion-examples/mce-cli/build/libs/mce-cli.jar -m produce
 ```
 
 Where `my-file.json` is some file that contains a
-[MetadataChangEvents](./src/main/pegasus/com/linkedin/metadata/examples/cli/MetadataChangeEvents.pdl) JSON object.
+[MetadataChangEvents](./src/main/pegasus/com/linkedin/metadata/examples/cli/MetadataChangeEvents.pdl) JSON object.
+
+### Producing the Example Events with Docker
+
+We have some example events in the `example-bootstrap.json` file, which can be invoked via the above example or in a
+docker environment using `docker/ingestion/ingestion.sh`. We also have a developer image
+(`docker/ingestion/ingestion-dev.sh`) which uses your locally built jar rather than building on the docker image itself,
+which may be faster if you have already built code locally.