Resource relationships in Managed Service for OpenSearch

OpenSearch is a specialized distributed system for storing, searching, visualizing, and analyzing data. When you add a document to OpenSearch, it is placed into an index. Data from an indexed document is available for search and analysis.

OpenSearch features OpenSearch Dashboards, a data visualization UI.

The main entity used in Managed Service for OpenSearch is a database cluster.

Each cluster consists of one or more hosts, which are virtual machines with DBMS servers deployed. Cluster hosts may reside in different availability zones. The minimum number of hosts in a cluster depends on the selected disk type.

When creating a cluster, specify:

• Host classes: VM templates for deploying the cluster hosts.

• Host groups: Groups of multiple linked hosts with specific configuration and roles:

  • DATA: Enable index storage and request processing.
  • MANAGER: Ensure fault tolerance of all OpenSearch components.
  • DASHBOARDS: Provide access to OpenSearch Dashboards.

  A single-host cluster is not fault tolerant. A cluster with multiple hosts automatically ensures fault tolerance for the MANAGER role. To ensure data safety in such a cluster, you need to configure index sharding and replication.

• Environment where the cluster will be deployed:

  • PRODUCTION: For stable versions of your apps.
  • PRESTABLE: For testing purposes. The prestable environment is similar to the production environment and likewise covered by the SLA, but it is the first to get new functionalities, improvements, and bug fixes. In the prestable environment, you can test compatibility of new versions with your application.

An OpenSearch cluster created in a folder can be accessed by all VMs connected to the same cloud network as the cluster.

Cluster fault toleranceCluster fault tolerance

For secure cluster data storage and continuous availability, the cluster must meet the following requirements:

• The cluster must not have any indexes with zero replicas or settings that prevent replicas from being created.
• Cluster hosts with the DATA role must reside in at least two different availability zones.
• Cluster hosts with the MANAGER role must reside in at least three different availability zones; and each availability zone must have less than 50% of such hosts of the total number of hosts with the MANAGER role.