Introducing time series collections for the next generation of Amazon OpenSearch Serverless

This article announces time series collections for the next generation of Amazon OpenSearch Serverless, enabling ingestion and analysis of logs, metrics, traces, and event data without cluster management.

• Automatically absorb ingest spikes and scale compute to zero when idle, eliminating worst-day capacity planning
• Hot tier stores recent data for fast queries; warm tier on S3 keeps years of history searchable at lower cost
• Configure data tiering with index.warm.after setting (1 day to 10 years) and retention independently
• Support for up to 100 TB hot storage per collection for high-scale observability workloads
• Built-in alerting and anomaly detection using Random Cut Forest machine learning algorithm
• Timestamp field required for all documents; configurable via index.timestamp_field setting
• Available in all AWS commercial regions where next-generation OpenSearch Serverless is available

Time series collections eliminate tradeoffs between cost, retention, and query speed for observability data management.