The Million-Tenant Problem: Why Your OpenSearch Service Architecture Breaks at Scale

Short summary

Multi-tenant OpenSearch architectures that work at hundreds of tenants break at hundreds of thousands due to fan-out query costs, cluster state bloat, and tiering latency mismatches. The author identifies three common traps: pure-logging approaches that scatter-gather across too many shards, pure-search approaches that create unmanageable index counts, and log-tiering strategies that add unacceptable latency for user-facing queries. Practical architectural patterns for million-tenant scale are introduced.

• •Pure-logging shard strategy causes query fan-out across all shards, swamping coordination costs at scale
• •Pure per-tenant indexing collapses under metadata management overhead beyond tens of thousands of tenants
• •Log-tiering approaches like UltraWarm add seconds of latency, unacceptable for user-facing search

Generated with AI, which can make mistakes.