Deploying an Elastic Stack - Overview

Scalingo for Elasticsearch® Multi-buildpack

The Elastic Stack (formerly known as the ELK Stack) is a powerful collection of softwares that lets you collect data from any source using any format. It gives you tools to search, visualize and analyze these data in real time.

The Elastick Stack is based on three major components: Elasticsearch®, Logstash and Kibana.

Here is a diagram showing the main principles of the stack architecture:

Planning your Deployment

  • Because most of the time, logs are only relevant for a short period of time, it is generally a good idea to archive or remove them after this short period to keep indices as light and fast as possible.

    To cover this need, we suggest to add a fourth component named Curator to your Elastic Stack.

  • Logstash and Kibana both require their own container(s). We will consequently need two apps.

  • We will deploy Elasticsearch as an addon attached to Logstash.

  • We will deploy Curator alongside Logstash, in the same container.

  • Logstash requires quite a lot of RAM to run properly. We recommend to deploy at least one L container to host it.

  • Choosing the appropriate Elasticsearch plan strongly depends on your needs. In this guide, we will start with the smallest plan we provide: a Starter 512M plan.

    Note that Elasticsearch can become quite greedy memory-wise: indexing, aggregating, and heavy read workloads with caching enabled can have a significant impact on your database performances. To reduce the JVM heap pressure, and to lower the need of garbage collection, you may want to switch for a more powerful plan.

Deploying

Please refer to our dedicated pages:

Suggest edits

©2026 Scalingo

GTC GTU DPA SLA HDS