Resize Your Cluster

Elasticsearch scales to whatever capacity you need and with as many nodes as the available resources can support. If you don’t have enough available resources, add some capacity first.

To resize a cluster:

  1. Log into the Cloud UI.
  2. Click on a cluster name from the Clusters panel and click Manage.
  3. Click Edit configuration.
  4. Change the cluster configuration:

    Fault tolerance

    If the initial cluster you created uses only one availability zone, it is not fault tolerant. On a production system, enable high availability by changing your cluster to use at least two availability zones, three for mission-critical deployments.

    Warning

    Clusters that use only one availability zone are not highly available and are at risk of data loss, if you do not configure an external snapshot repository to enable regular backups. To safeguard against data loss, you must use at least two data centers and configure an external repository for backups.

    Node Capacity
    Node capacity should be sufficient to sustain your search workload, even if you lose an availability zone. Currently, half of the memory is assigned to the JVM heap. For example, on a cluster with 32 GB RAM, 16 GB are allotted to heap. Up to 64 GB RAM and 1 TB storage per node are supported.
    Node Count
    Adding more nodes lets you scale out horizontally by adding more processing capacity to your cluster.
  5. Click Save changes.

Example: From Very Small to Very Large

The first cluster we provisioned was very basic: It used only one availability zone and a singe node. In this example, you change the same cluster to use high availability and to add capacity.

To scale your cluster from very small to very large:

  1. Log into the Cloud UI.
  2. Click on a cluster name from the Clusters panel and click Manage.
  3. Click Edit configuration.
  4. Under Fault tolerance, select 3 zones — For mission critical environments.
  5. Under Node capacity, select 64 GB memory / 2 TB storage.
  6. Click Save changes.

There is no downtime when adding high availability or increasing cluster capacity. Your cluster continues to handle user requests as the configuration change gets applied.