The next infrastructure service is Elasticsearch, which also has no dependencies on other services. It will be used by the message handler, which also uses the NATS message queue, so I will need to join all these services to the same Docker network. For Elasticsearch, I also want to limit the amount of memory it uses and use a volume for the data so it will be stored outside the container:

 elasticsearch:
   image: sixeyed/elasticsearch:5.6.11-windowsservercore-ltsc2019
   environment: 
     - ES_JAVA_OPTS=-Xms512m -Xmx512m
   volumes:
     - es-data:C:data
   networks:
     - nd-net

Here, elasticsearch is the name of the service and sixeyed/elasticsearch is the name of the image, which is my public image on Docker Hub. I'm connecting the service to the same nd-net network, and I also mount a volume to a known location in the container. When Elasticsearch writes data to C:data on the container, it will actually be stored in a volume.

Just as with networks, volumes are first-class resources in the Docker Compose file. For Elasticsearch, I'm mapping a volume called es-data to the data location in the container. I'll specify how the es-data volume should be created later in the Compose file.