Amazon Managed Service for Prometheus Supports Out-of-Order Metric Ingestion

New features enable ingestion of late-arriving metric samples and rule query delay in Amazon Managed Service for Prometheus.

By Eva Hart · Editorial pseudonym · 2026-06-12 · 1 min read

AI-ASSISTED REPORTING · RETROSPECTIVE ARCHIVE

Retrospective edition: 2026-06-12. Source published: 2026-06-11. Added to this archive: 2026-11-03. This briefing was prepared later from the linked announcement; it was not published on this website on the historical edition date.

By default, a one-minute out-of-order time window lets Prometheus ingests metric samples arriving out of strict chronological order, preventing data loss in distributed or variable-latency environments. Users can adjust this window or disable it if desired.

Additionally, a workspace-level rule query offset delays rule evaluation queries, allowing late samples to be ingested before alerts run. Together these innovations preserve metric completeness and improve alert accuracy. Monitoring via new CloudWatch metrics helps optimize configuration.

Source & further reading

Amazon Managed Service for Prometheus now supports out of order sample ingestion ↗

Source published 2026-06-11. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.