Loading

Upgrade

Stack ECE ECK ECH Self-Managed

Upgrading to the latest version provides access to the newest Elastic features, security patches, performance improvements, and bug fixes. These updates reduce costs, speed up threat response, and improve investigative and analytical data tools.

As Elastic releases new versions, older versions of Elastic products reach their end of life on a set schedule. To keep your deployment supported, stay up to date. For more information, refer to Product End of Life Dates.

Note

With Elastic Cloud Serverless, upgrades of the Elastic-managed infrastructure and project components are fully handled by Elastic. Users automatically receive the latest features and improvements, with no need to plan or perform project-level upgrades.

Keep in mind that the ingest components or clients you manage externally, such as Elastic Agent, Beats, Logstash, or custom applications, still need to be updated manually.

The upgrade procedure depends on how your deployment is managed. If you're using Elastic Cloud Hosted or Elastic Cloud Enterprise, upgrades can often be performed with a single click in the Elastic Cloud UI. For self-managed deployments, upgrades must be carried out manually using a rolling upgrade process, upgrading the nodes one by one to minimize downtime and ensure cluster stability.

This section provides guidance to help you plan and safely perform upgrades of your Elastic Stack components, with a primary focus on Elasticsearch and Kibana as the core of the stack. It also covers upgrades for orchestration platforms like Elastic Cloud Enterprise and Elastic Cloud on Kubernetes, as well as related components such as APM, Beats, Elastic Agent, and Logstash.

Important

In Elastic Stack 9.0.0 and beyond, Enterprise Search is unavailable. For more information, refer to Migrating to 9.x from Enterprise Search 8.x versions.

Upgrading your Elastic cluster or deployment involves several stages, including planning, preparation, and execution. This section guides you through the full upgrade process:

Additionally, if you're using a self-managed orchestration platform such as Elastic Cloud Enterprise or Elastic Cloud on Kubernetes, refer to Upgrade your ECE or ECK orchestrator to keep the orchestrator up to date.

Elasticsearch supports rolling upgrades, where nodes are upgraded one at a time to minimize downtime and ensure cluster stability. If your deployment or cluster follows high availability best practices, such as using replicated indices and distributing nodes across multiple availability zones, it should remain available throughout the entire upgrade process.

Kibana upgrades, however, require downtime. All running instances must be shut down before starting the upgrade.

Important

Ensure your deployment is properly sized and configured for high availability to reduce the risk of service interruption during upgrades.

If you want to learn more about scaling, high availability, and performance, refer to:

You’ll find more information about availability considerations for specific deployment types in the related upgrade guides within this section.

You can upgrade to a higher version if the target version was released after your current version. Upgrades to versions released before your current version are not supported, even if the version number is higher. Refer to out-of-order releases for more information.

For example:

  • ✅ Upgrade allowed: From 9.0.4 to 9.1.0 (9.1.0 released after 9.0.4)
  • ❌ Not allowed: From 9.0.5 to 9.1.0 (9.1.0 released before 9.0.5) → wait for 9.1.1 to be released

To perform a major upgrade from 8.x, the required starting version depends on the target release:

  • To upgrade to the 9.0.x series, you must be on 8.18.x.
  • To upgrade to 9.1.0 or later, you must be on 8.19.x, which is the latest minor release of the 8.x series.
Note

While 8.19 is the final minor release in the 8.x series, 8.18 was released at the same time as 9.0, enabling a supported upgrade path between the 8.18.x and 9.0.x series. This compatibility also applies to other features and clients.

If you're upgrading to the current 9.1.0 release from an earlier 8.x version, first upgrade to the latest available 8.19 release.

For flexible upgrade scheduling, 8.19 Elastic Agent, Beats, and Logstash are compatible with all 9.x versions of Elasticsearch.

By default, 8.x Elasticsearch clients are compatible with 9.x and use REST API compatibility to maintain compatibility with the 9.x cluster.