AOSCOST01-BP03 Evaluate manager nodes - Amazon OpenSearch Service Lens

AOSCOST01-BP03 Evaluate manager nodes

Improve domain stability, reduce risk of data loss, and prevent over-provisioning by using an appropriate instance type for dedicated leader nodes.

Level of risk exposed if this best practice is not established: High

Desired outcome: The correct instance type is used for dedicated leader nodes in OpenSearch Service to gain optimal resource utilization and performance.

Benefits of establishing this best practice:

  • Enhanced domain stability, as the leader role is handled by a dedicated node. This practice avoids unnecessary resource utilization and potential over-provisioning of data nodes.

  • Reduced risk of data loss or corruption due to inadequate resource allocation or leader failure.

Implementation guidance

Consider dedicated manager nodes size correlated with data node instance sizes, number of instances, indexes, and shards they can manage. For production clusters, we recommend, at a minimum, the following instance types for dedicated leading nodes:

Instance count Manager node RAM size Maximum supported shard count Recommended minimum dedicated leader instance type
1–10 8 GiB 10K m5.large.search or m6g.large.search
11–30 16 GiB 30K c5.2xlarge.search or c6g.2xlarge.search
31–75 32 GiB 40K r5.xlarge.search or r6g.xlarge.search
76–125 64 GiB 75K r5.2xlarge.search or r6g.2xlarge.search
126–200 128 GiB 75K r5.4xlarge.search or r6g.4xlarge.search

Resources