Elastic Stack Installation Overview

This page provides a high-level overview of the Elastic Stack installation process and helps identify the correct installation path for the environment.

A Relativity Elastic Stack deployment includes Elasticsearch, Kibana, and APM Server (when applicable). The required Elastic Stack components depend on environment configuration and system requirements. For details, see the Elastic Stack System Requirements.

For information about supported Elasticsearch versions and deployment approaches for existing Data Grid clusters, see Supported Elastic Stack Versions.

Cluster Architecture Decision

Before beginning installation, you must decide on your cluster architecture:

Option 1: Single Unified Cluster (single cluster with multiple nodes)
  • One Elasticsearch cluster that contains all data (Environment Watch and Data Grid Audit)
  • Simpler to manage and maintain
  • Shared resources and infrastructure
  • Suitable for most deployments
Option 2: Separate Clusters (multi-cluster)
  • Two independent Elasticsearch clusters:
    • One dedicated cluster for Environment Watch
    • One dedicated cluster for Data Grid Audit
  • Complete isolation between workloads
  • Independent scaling and resource allocation
  • Recommended for very large deployments or when strict data separation is required

For most organizations, Option 1 (Single Unified Cluster) is recommended as it simplifies operations while providing adequate performance and isolation through index management.

Offline Environment Considerations

For Elastic's guidance on air-gapped installations, see Air-gapped install.

If your environment is offline or air-gapped (no direct internet access), you have two options for obtaining Elastic Stack installation files:

Option 1: Proxy Server (Recommended)

Using a proxy server is the simplest approach for air-gapped environments:

  1. Configure a proxy server that has internet access and is reachable from your isolated network.

  2. Set proxy environment variables on the target servers before downloading:

    Copy 
    $env:HTTP_PROXY = "http://proxy.yourcompany.com:8080"
    $env:HTTPS_PROXY = "http://proxy.yourcompany.com:8080"

  3. Download files normally using the URLs in this guide — traffic will route through the proxy.

Benefits:

  • Minimal workflow changes from standard installation
  • Automatic dependency resolution for plugins
  • Easier to maintain and update over time
Option 2: Manual File Transfer (Air-Gapped)

For fully isolated networks with no proxy access:

  1. From an internet-connected machine, download all required files:

    • Elasticsearch ZIP (e.g., elasticsearch-8.x.x-windows-x86_64.zip)
    • Kibana ZIP (e.g., kibana-8.x.x-windows-x86_64.zip)
    • APM Server ZIP (e.g., apm-server-8.x.x-windows-x86_64.zip)
    • RWSM NuGet package from https://relativitypackageseastus.jfrog.io/artifactory/api/nuget/server-nuget-virtual/Download/Relativity.Windows.ServiceManager/2.24.0
    • Any required plugins from Elastic's plugin repository

  2. Transfer files to the air-gapped environment using approved media (USB drive, secure file transfer, etc.).

  3. Install plugins offline using local file paths:

    Copy 
    .\elasticsearch-plugin install file:///C:/downloads/mapper-size-8.x.x.zip

Considerations:

  • Requires more planning and coordination
  • All dependencies must be manually tracked and transferred
  • Plugin versions must exactly match the Elasticsearch version

Installation Paths

Once you have reviewed the considerations above, choose the appropriate installation method for your environment:

| Automated and Manual

Return to top of the page
Feedback