Customers are now able to scale their IBM Cloud Databases for Elasticsearch instances up to 20 members.
This new feature allows Elasticsearch instances to now grow to 2.2TB RAM, 80 TB of Disk, and 560 vCPUs.
Notes on scaling
Keep in mind the following information as you scale your Elasticsearch usage:
- A higher number of nodes provides higher availability in the case of node failures. As long as (n/2)+1 nodes are online, the cluster can elect a master node and accept write requests. So, in a 3-node cluster, 2 lost nodes stop the database. But a 10-node cluster can lose 4 nodes before that happens.
- Make sure to adjust the replica setting for your indices if you want more replicas in a larger cluster. Consider using the
auto_expand_replicasfeature if you plan to increase the cluster frequently.
- Nodes can only be scaled out at this time—you are not able to decrease the amount of servers allocated to Databases for Elasticsearch.
You can review our Horizontal Scaling documentation for more information and the relevant API endpoints.
IBM Cloud Databases for Elasticsearch is an enterprise-ready, fully-managed Elasticsearch, built with native integration into the IBM Cloud.