This is a little follow up to a post I did awhile back about getting the ELK stack up and running using Docker. The last post was over a year ago and a lot has changed in regards to both Docker and the ELK stack.
All of the components of the ELK stack have gone through several revisions since the last post and all kinds of features and improvements have been made to all components (Elasticsearch, Logstash and Kibana). The current iteration is v5 for all of the components. v5 is still in alpha but that doesn’t mean we can’t get it up and running with Docker. NOTE: I don’t recommend trying to run ELK v5 in any kind of a setup outside of development at this point since it is still alpha.
Docker has evolved a little bit as well since the last post, which will help with some of the setup. The improvements in docker-compose will allow us to wrap the new Docker features up in the containers and leverage some cool Docker features.
Here is the updated elk-docker repo. Please open a PR or issue if you have ideas for improvement or if there are any issues you run into.
For the most part the items in the repo shouldn’t need to change unless you are interested in adjusting the Elasticsearch configuration or you want to update the Logstash input/filter/output configuration. The Elasticsearch config is located in es/elasticsearch.yml and the Logstash config is located in logstash/logstash.conf.
This configuration has been tested using Docker version 1.11 and docker-compose 1.7 on OS X.
Here’s what the docker-compose file looks like.
version: '2' services: elasticsearch: image: elasticsearch:5 command: elasticsearch environment: # This helps ES out with memory usage - ES_JAVA_OPTS=-Xmx1g -Xms1g volumes: # Persist elasticsearch data to a volume - elasticsearch:/usr/share/elasticsearch/data # Extra ES configuration options - ./es/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml ports: - "9200:9200" - "9300:9300" logstash: image: logstash:5 command: logstash --auto-reload -w 4 -f /etc/logstash/conf.d/logstash.conf environment: # This helps Logstash out if it gets too busy - LS_HEAP_SIZE=2048m volumes: # volume mount the logstash config - ./logstash/logstash.conf:/etc/logstash/conf.d/logstash.conf ports: # Default GELF port - "12201:12201/udp" # Default UDP port - "5000:5000/udp" # Default TCP port - "5001:5001" links: - elasticsearch kibana: image: kibana:5 environment: # Point Kibana to the elasticsearch container - ELASTICSEARCH_URL=http://elasticsearch:9200 ports: - "5601:5601" links: - elasticsearch kopf: image: rancher/kopf:v0.4.0 ports: - "8080:80" environment: KOPF_ES_SERVERS: "elasticsearch:9200" links: - elasticsearch volumes: elasticsearch:
Notice that we are just storing the Elasticsearch data in a Docker volume called “elasticsearch”. Storing the data in a volume makes it easier to manage.
To start up the ELK stack just run docker-compose up” (plus -d for detatched) and you should see the ELK components start to come up in the docker-compose log messages. It takes about a minute or so to come up.
After everything has bootstrapped and come up you can see the fruits of your labor. If you are using the Docker beta app, (which I highly recommend) you can just visit localhost:5601 in your browser.
To easily get some logs into ELK to start playing around with some data you can run the logspout container like I have below.
docker run --rm --name="logspout" \ --volume=/var/run/docker.sock:/var/run/docker.sock \ --publish=127.0.0.1:8000:80 \ gliderlabs/logspout:master \ syslog://<local_ip_address>:5001
The value of <local_ip_address> should be the address of your laptop or desktop, which you can grab with ifconfig.
That’s pretty much all there is to it. Feel free to tweak the configs if you want to play around with logstash or elasticsearch. And also please let me know if you have any ideas for improvement or have any issues getting this up and running.