Unconference Session: Dedicated Coordination Nodes - Why? When? How?
summary
Amazon OpenSearch Service adds dedicated coordination nodes (DCN) to isolate coordination work like shard routing, aggregations, hybrid/vector normalization and exporting millions of doc IDs. Use DCNs for 50–100 data nodes, high-QPS e-commerce clusters, heavy aggregations or deep pagination to scale without reindexing. Sizing: CPU and JVM heap spike on coordinators while off‑heap/file cache stays low; choose CPU-optimized instances for indexing-heavy work or slightly larger memory instances for search-heavy work (example: R8G 2x → R8G 4x), monitor CPU >80%, run at least two DCNs and place DCNs on targets for cross-cluster search.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:02Coordination role in search
- 3:29Resource patterns in coordinator
- 4:36When to use coordinator nodes
- 6:00Scaling strategies
- 7:43Underutilized metrics analysis
- 9:10Sizing coordinator nodes
- 12:43Number of coordinator nodes
- 13:28Cross-cluster search & failover
- 13:49Q&A
related talks
44mBringing Down the Cost of Your OpenSearch Clusters
31mMaximize Resource Efficiency with Separated Index and Search Workloads
41mDynamoDB, Neptune and OpenSearch: A Unified Approach to Collaborative Editing and Search #JSConfAM25
44mScaling OpenSearch at Uber: Operating Resilient, Fault-Tolerant, High-Performance Clus...
45mMonitoring Your OpenSearch Cluster Effectively
43mMulti-tenancy for all workloads
39mScaling OpenSearch: Next-Generation Shard Allo...
20mLightning Talk: Multi-Datacenter Strategies for Data Services