The application is running in Docker using docker-compose. The application itself consists of component1 and component2 illustrating 2 Mule applications with top level flow in component1 calling a number of flows in component2 over http. Docker-compose brings up the Elastic stack (Kibana, Elasticsearch and APM server) that are wired to receive traces from component1 and component2. The application also contains a steady state load generator invoking component1 at about 10 requests per second rate to create a stable base load.
Ensure docker daemon is configured with 8GB RAM. Also, if you use proxies, you may need to update Dockerfile to copy across your Maven settings.xml file.
Before build, set the following environment variables, if building on Mac:
export DOCKER_BUILDKIT=0
export COMPOSE_DOCKER_CLI_BUILD=0
In directory example-app, start the application by running TAG=6.6.1 docker-compose up that includes the desired version of the ELK stack. The application will download a whole lot of maven dependencies in the first run, so give it some time. Once up, you will see the following log entries telling that it is fully up:
apm-server-apm-mule | 2019-02-01T12:39:26.995Z INFO [request] beater/common_handlers.go:272 handled request {"request_id": "2eed0618-6b6f-4c5a-9c39-6572c6bd810b", "method": "POST", "URL": "/intake/v2/events", "content_length": -1, "remote_address": "172.23.0.6", "user-agent": "java-agent/1.3.0", "response_code": 202}
Once the docker containers are up, navigate to http://localhost:5601 and select APM from the left-hand menu. For the first run, go through Setup and select Load Kibana Objects.
To trigger transactions, send a GET request to http://localhost:8281. It should return 200 Ok status and success as a body of the response. You can use curl to use it from command line, or click on this link:
curl -v http://localhost:8281
To stop, press ^C in the terminal with docker-compose. To remove all the created containers, run:
docker container prune -f