What it does
Amazon Managed Service for Prometheus is an AWS-managed service for storing, querying, and alerting on Prometheus metrics, especially for containerized workloads. It accepts Prometheus remote-write data and can use an AWS-managed collector to scrape EKS metrics or a collector that the customer operates. Teams use PromQL and Prometheus-compatible tooling while AWS manages the service infrastructure. It is a metrics backend, not a replacement for the wider logs, traces, or application-error tools in an observability stack.
Create a workspace, then connect a collector
The AWS getting-started flow has three main stages: create a workspace, ingest metrics, and query them. For EKS, the service can scrape compatible endpoints with an AWS-managed collector; teams can also run their own Prometheus agent or AWS Distro for OpenTelemetry collector and send data through Prometheus remote write. Once metrics arrive, use PromQL to explore the data and create alerts or connect a Prometheus-compatible dashboard.
A useful starting task is to choose one Kubernetes workload, confirm that its metrics are available at a Prometheus-compatible endpoint, and send that stream to a workspace. Keep the workspace focused on metrics that support a specific operational question, such as pod saturation or request rates, before broadening collection. This is a managed metrics store, so diagnosing an application error may still require logs, traces, or another tool.
AWS documents a 150-day default retention period for workspace data, configurable up to 1,095 days. Metrics are deleted after the configured retention period. An AWS account is needed to create the workspace, and ingestion/query setup is part of the service. The official pricing page was accessed, but its current rate table was not exposed in the page text retrieved for this research; no billing classification or exact cost is inferred from that incomplete view.
Strengths
- Uses Prometheus data and PromQL while AWS manages the service's ingestion, storage, and query infrastructure.
- Can scrape EKS endpoints with an AWS-managed collector or accept metrics from customer-managed Prometheus or OpenTelemetry collectors.
- Supports container metrics from EKS and self-managed Kubernetes environments.
Watch out for
- The service is centered on Prometheus metrics for container workloads, not a full logs-and-traces observability suite.
- Workspace data expires after the configured retention period; the default is 150 days unless an operator changes it.
Best for
- Kubernetes teams that want managed Prometheus-compatible metrics storage and PromQL queries
- EKS operators seeking a managed scraper instead of maintaining their own collector
- AWS users connecting a Prometheus metrics backend to existing dashboards and alerts
Sources
- Container Environment Monitoring Service - Amazon Managed Service for Prometheus (Oct 26, 2026)
- What is Amazon Managed Service for Prometheus? (Oct 26, 2026)
- Get started with Amazon Managed Service for Prometheus (Oct 26, 2026)
- Ingest metrics to your Amazon Managed Service for Prometheus workspace (Oct 26, 2026)
- Create an Amazon Managed Service for Prometheus workspace (Oct 26, 2026)
- Amazon Managed Service for Prometheus Pricing (Oct 26, 2026)