An opinionated Docker implementation for Elasticsearch focused on providing a reliable yet flexible solution based on the upstream official image.
- Features
- Getting Started
- Deployment
- Components
- Contributing
- Versioning
- Authors
- License
- Acknowledgments
There are a number of features that differentiate this implementation from the official or other solutions.
- Configuration determined at runtime by a powerful templating engine.
- AWS availability zone awareness auto detection.
- Extra configuration snippets easy to add.
- Heap sized based on container memory and percentages.
- Enables AWS S3 plugin for snapshots.
- Enables Prometheus plugin for metrics.
This implementation is designed to be used with a Helm chart. However, there is a docker compose file included for testing this implementation locally. Bringing up a test cluster and discovering the client endpoint is easily done with just 2 commands.
docker-compose up
It will likely take a few minutes for the containers to start and you can verify when it is available using curl.
curl $(docker-compose port client 9200)
The only requirements for bringing up this implementation locally is Docker.
Docker 17.12.0+
However, be aware that Elasticsearch has significant memory requirements especially since it is Java based and it is necessary to allocate the heap on start to prevent memory fragmentation. With 4 Elasticsearch containers (Discovery, Master, Data and Client), Kibana and Cerebro at least 4388Mi of RAM is required in heap allocation, preferably 6Gi if there is capacity.
It is expected that the Helm chart will be used to install this container. For local development, there are a number of ways to interact with Elasticsearch.
The client endpoint is the main way applications interact with Elasticsearch.
docker-compose port client 9200
Management of Elasticsearch is available using Kibana.
docker-compose port kibana 5601
An alternative management interface is available using Cerebro.
docker-compose port cerebro 9000
| Variable | Default | Comment |
|---|---|---|
| ES_CLUSTER_NAME | example | |
| ES_NETWORK_HOST | 0.0.0.0 | |
| ES_NODE_NAME | node | |
| ES_NODE_MASTER | false | |
| ES_NODE_DATA | false | |
| ES_XPACK_SECURITY_ENABLED | false | |
| ES_XPACK_MONITORING_ENABLED | true | |
| ES_XPACK_MONITORING_COLLECTION_ENABLED | true | |
| ES_EMAIL_AWS_SES_ENABLED | false | |
| ES_EMAIL_AWS_SES_HOST | email-smtp.us-east-1.amazonaws.com | |
| ES_EMAIL_AWS_SES_USER | Unset | |
| ES_EMAIL_AWS_SES_PASSWORD | Unset | |
| ES_HTTP_CORS_ENABLED | false | |
| ES_HTTP_CORS_ALLOW_ORIGIN | * | |
| ES_GATEWAY_RECOVER_AFTER_DATA_NODES | 1 | |
| ES_GATEWAY_EXPECTED_DATA_NODES | 1 | |
| ES_GATEWAY_RECOVER_AFTER_TIME | 1m | |
| ES_DISCOVERY_ZEN_MINIMUM_MASTER_NODES | 1 | |
| ES_DISCOVERY_ZEN_PING_UNICAST_HOSTS | localhost | |
| ES_DISCOVERY_TYPE | Unset |
The existence of /usr/share/elasticsearch/config/elasticsearch_custom.yml will automatically be appended to the elasticsearch.yml settings at runtime.
Running a JVM in a container requires an understand of the relationship of the cgroups memory limit and the heap size. There is an overhead required in running a JVM as well which is approximate 348Mi.
A simple formula to use to determine the RAM percentage is as follows:
(Memory Limit - 384Mi / Memory Limit) * 100
For use on live systems, see the documentation for the Helm chart.
- Elasticsearch - Open Source, Distributed, RESTful Search Engine
- Elasticsearch-Docker - Official Elasticsearch Docker image from Elastic.
- Gomplate - A flexible commandline tool for template rendering. Supports lots of local and remote datasources.
- Kibana - Kibana analytics and search dashboard for Elasticsearch
- Cerebro - Cerebro is an open source Elasticsearch web admin tool built using Scala, Play Framework, AngularJS and Bootstrap
- Prometheus Exporter Plugin for Elasticsearch - Prometheus exporter plugin for Elasticsearch
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.
We use SemVer for versioning. For the versions available, see the tags on this repository.
- Michael Lorant - Initial work - Nine
See also the list of contributors who participated in this project.
This project is licensed under the MIT License - see the LICENSE.md file for details
- Elastic
- Nine