# Amazon Managed Service for Prometheus

Cloud service · AWS. Updated 2026-11-12. Page: https://blowstack.com/tools/cloud-services-hub/amazon-managed-service-for-prometheus

> Amazon Managed Service for Prometheus is a fully managed Prometheus-compatible monitoring service that makes it easy to monitor containerized applications at scale on AWS.

Practice exams covering Amazon Managed Service for Prometheus: [SOA-C03 AWS Certified CloudOps Engineer - Associate](https://blowstack.com/practice-exams/aws-certified-cloudops-engineer-associate), [DOP-C02 AWS Certified DevOps Engineer - Professional](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional), [MLA-C02 (beta) AWS Certified Machine Learning Engineer - Associate](https://blowstack.com/practice-exams/aws-certified-machine-learning-engineer-associate), [SAP-C03 AWS Certified Solutions Architect - Professional](https://blowstack.com/practice-exams/aws-certified-solutions-architect-professional).

## 6 key features of Amazon Managed Service for Prometheus

Amazon Managed Service for Prometheus is a secure, scalable, and fully managed monitoring service that offers serverless operation, native PromQL support, secure data management, and seamless integration with the AWS ecosystem.

1. **Secure and Scalable Metrics Collection**: Amazon Managed Service for Prometheus offers a fully managed, secure, and scalable environment for collecting, querying, and storing Prometheus metrics. It automatically scales as your monitoring workloads grow, ensuring that your metrics collection can handle the demands of large-scale applications with ease.
2. **Serverless Operational Mode**: With a serverless operational mode, users do not need to manage the underlying infrastructure. This service abstracts the complexity of running and operating Prometheus at scale, reducing the operational overhead and making it easier for users to focus on their core application development.
3. **Native Prometheus Query Language Support**: The service provides full support for Prometheus Query Language (PromQL), enabling users to run their existing queries without modifications. This seamless integration allows developers to leverage their existing Prometheus dashboards and alerts, facilitating a smooth migration to the managed service.
4. **Secure Data Storage and Management**: Amazon Managed Service for Prometheus ensures that your monitoring data is securely stored and managed. It provides automatic encryption of your data at rest and in transit, along with fine-grained access control mechanisms to manage access to your Prometheus metrics.
5. **Integrated with AWS Ecosystem**: This service is deeply integrated with the AWS ecosystem, providing a unified experience when monitoring AWS resources.
6. **Multi-Availability Zone Replication**: Data ingested into a workspace is replicated across three Availability Zones in the same Region for high availability.

## Where Amazon Managed Service for Prometheus is used

1. **Real-Time Monitoring of Kubernetes Clusters**: Using Amazon Managed Service for Prometheus, you can set up real-time monitoring for your Kubernetes clusters. This allows you to collect, store, and query metrics about the cluster's performance and the state of applications running on it. Enhanced metrics collection and querying capabilities enable you to quickly identify and address performance bottlenecks, ensuring high availability and reliability of your services.
2. **Infrastructure Monitoring Across AWS Services**: Leverage Amazon Managed Service for Prometheus to monitor the health and performance of various AWS services in your infrastructure. By integrating with AWS CloudWatch, you can aggregate metrics from different sources, providing a comprehensive view of your entire infrastructure. This holistic approach aids in proactive issue resolution and efficient resource utilization.

## How Amazon Managed Service for Prometheus is priced

Amazon Managed Service for Prometheus pricing includes charges for stored compressed metric data, data transfers, and the number of active series.

- **On-Demand Metrics Storage**: Storage is billed for metric samples and metric metadata in gigabytes, with 1 GB equal to 2³⁰ bytes; metadata size depends on metric names and associated labels.
- **Data Transfer**: Charges apply for data transferred out of Amazon Managed Service for Prometheus to the internet or other AWS regions, whereas data transfer in and between AWS services within the same region is generally free.
- **Metric Sample Ingestion**: Ingestion is billed by the number of metric samples ingested, with volume tiers based on the total number of samples ingested per month.
- **Query Samples Processed**: Queries are billed by Query Samples Processed (QSP), the total number of data points queried using PromQL for a given time range.
- **Managed Collector**: The Amazon Managed Service for Prometheus collector is billed for the number of hours it is turned on and the number of samples collected.

## Services Amazon Managed Service for Prometheus works with

- **[Amazon CloudWatch](https://blowstack.com/tools/cloud-services-hub/amazon-cloudwatch)**: Amazon Managed Service for Prometheus publishes workspace usage and operational metrics to CloudWatch and can send diagnostic logs there for troubleshooting. It can also ingest CloudWatch metrics delivered through Amazon Data Firehose, making them available for PromQL queries.
- **[Amazon Data Firehose](https://blowstack.com/tools/cloud-services-hub/amazon-kinesis-data-firehose)**: Amazon Managed Service for Prometheus can receive CloudWatch metric streams through an Amazon Data Firehose delivery stream configured to deliver to a workspace. This lets you query streamed AWS resource metrics alongside other Prometheus metrics.
- **[Amazon EC2](https://blowstack.com/tools/cloud-services-hub/amazon-ec2)**: Amazon Managed Service for Prometheus accepts metrics from Prometheus servers and AWS Distro for OpenTelemetry collectors running on EC2 instances. Collectors can scrape instance and application exporters and send the metrics to a workspace using authenticated remote write.
- **[Amazon ECS Anywhere](https://blowstack.com/tools/cloud-services-hub/amazon-ecs-anywhere)**: Amazon Managed Service for Prometheus can receive metrics from Amazon ECS Anywhere workloads through an AWS Distro for OpenTelemetry collector deployed with the workloads. This supports centralized monitoring of containers running on external hosts.
- **[Amazon EKS Anywhere](https://blowstack.com/tools/cloud-services-hub/amazon-eks-anywhere)**: Amazon Managed Service for Prometheus can store metrics collected from Amazon EKS Anywhere clusters using Prometheus remote write or an AWS Distro for OpenTelemetry collector. This allows on-premises Kubernetes metrics to be queried in a managed workspace.
- **[Amazon Elastic Container Service (Amazon ECS)](https://blowstack.com/tools/cloud-services-hub/amazon-elastic-container-service-(amazon-ecs))**: Amazon Managed Service for Prometheus integrates with Amazon ECS through AWS Distro for OpenTelemetry collectors deployed as ECS tasks or sidecars. The collectors scrape Prometheus-compatible application endpoints and remotely write their metrics to a workspace.
- **[Amazon Elastic Kubernetes Service (Amazon EKS)](https://blowstack.com/tools/cloud-services-hub/amazon-elastic-kubernetes-service-(amazon-eks))**: Amazon Managed Service for Prometheus provides managed collectors that discover and scrape Prometheus-compatible metrics from Amazon EKS clusters. EKS workloads can also send metrics through self-managed Prometheus or AWS Distro for OpenTelemetry collectors, including metrics used for Kubecost cost monitoring.
- **[Amazon EMR](https://blowstack.com/tools/cloud-services-hub/amazon-emr)**: Amazon Managed Service for Prometheus can receive Spark and JVM metrics from Amazon EMR on EKS through Prometheus exporters and an AWS Distro for OpenTelemetry collector. These metrics support monitoring job execution and resource use with PromQL and Grafana dashboards.
- **[Amazon Managed Grafana](https://blowstack.com/tools/cloud-services-hub/amazon-managed-grafana)**: Amazon Managed Service for Prometheus workspaces can be configured as data sources in Amazon Managed Grafana, which discovers workspaces and handles AWS authentication for queries. Grafana can visualize workspace metrics and display alerts from the workspace's alert manager.
- **[Amazon Managed Streaming for Apache Kafka](https://blowstack.com/tools/cloud-services-hub/amazon-managed-streaming-for-apache-kafka)**: Amazon Managed Service for Prometheus provides managed collectors for scraping the Prometheus-compatible metrics exposed by Amazon MSK open monitoring. The collected broker metrics can be stored in a workspace for querying and alerting.
- **[Amazon OpenSearch Service](https://blowstack.com/tools/cloud-services-hub/amazon-opensearch-service)**: Amazon Managed Service for Prometheus provides managed collectors that automatically scrape Prometheus-compatible metrics from Amazon OpenSearch Service domains. The collectors can forward those metrics to a Prometheus workspace or Amazon CloudWatch without requiring you to operate scraper instances.
- **[Amazon SageMaker](https://blowstack.com/tools/cloud-services-hub/amazon-sagemaker)**: Amazon Managed Service for Prometheus serves as a metrics backend for Amazon SageMaker HyperPod observability. Collectors on HyperPod clusters send infrastructure and accelerator metrics to a workspace for monitoring training-cluster health and utilization.
- **[Amazon Simple Notification Service (Amazon SNS)](https://blowstack.com/tools/cloud-services-hub/amazon-simple-notification-service-(amazon-sns))**: Amazon Managed Service for Prometheus uses Amazon SNS as an alert receiver for its managed alert manager. Alert routing rules can publish notifications to SNS topics for delivery to subscribers.
- **[Amazon VPC](https://blowstack.com/tools/cloud-services-hub/amazon-vpc)**: Amazon Managed Service for Prometheus managed collectors use network interfaces in your VPC to reach private metric endpoints, including endpoints in EKS clusters. VPC interface endpoints also let clients access supported service APIs without traversing the public internet.
- **[AWS CloudFormation](https://blowstack.com/tools/cloud-services-hub/aws-cloudformation)**: Amazon Managed Service for Prometheus supports CloudFormation resource types for workspaces, rule-group namespaces, and managed scrapers. Templates can provision these resources and maintain their configuration alongside the infrastructure they monitor.
- **[AWS CloudTrail](https://blowstack.com/tools/cloud-services-hub/aws-cloudtrail)**: Amazon Managed Service for Prometheus integrates with CloudTrail to record API activity. These records identify callers and requested operations, supporting audits and investigations of workspace administration and access.
- **[AWS Config](https://blowstack.com/tools/cloud-services-hub/aws-config)**: Amazon Managed Service for Prometheus workspaces are supported AWS Config resources. Config can record their configuration history so you can inspect changes and evaluate them against compliance requirements.
- **[AWS Distro for OpenTelemetry](https://blowstack.com/tools/cloud-services-hub/aws-distro-for-opentelemetry)**: Amazon Managed Service for Prometheus accepts metrics through the AWS Distro for OpenTelemetry collector's Prometheus remote-write exporter. The collector can scrape application endpoints, authenticate requests with AWS credentials, and send metrics to a workspace.
- **[AWS Fargate](https://blowstack.com/tools/cloud-services-hub/aws-fargate)**: Amazon Managed Service for Prometheus can monitor applications running on AWS Fargate through AWS Distro for OpenTelemetry collectors deployed with ECS tasks or EKS pods. The collectors forward application metrics to a workspace without requiring access to the underlying hosts.
- **[AWS Identity and Access Management (IAM)](https://blowstack.com/tools/cloud-services-hub/aws-identity-and-access-management-(iam))**: Amazon Managed Service for Prometheus uses IAM policies and roles to authorize workspace administration, metric ingestion, and metric queries. Separate roles can grant collectors permission to write metrics while allowing dashboard users only to query them.
- **[AWS Key Management Service (AWS KMS)](https://blowstack.com/tools/cloud-services-hub/aws-key-management-service-(aws-kms))**: Amazon Managed Service for Prometheus supports customer-managed AWS KMS keys for encrypting workspace data at rest. Key policies and grants control the service's use of the key.
- **[AWS Managed Services (AMS)](https://blowstack.com/tools/cloud-services-hub/aws-managed-services-(ams))**: Amazon Managed Service for Prometheus can be provisioned in an AMS-managed account through AMS Self-Service Provisioning. This allows you to use managed Prometheus workspaces within the account's AMS operating environment.
- **[AWS PrivateLink](https://blowstack.com/tools/cloud-services-hub/aws-privatelink)**: Amazon Managed Service for Prometheus supports private access through AWS PrivateLink interface endpoints. Collectors and query clients can use supported service endpoints from a VPC without an internet gateway or public IP addresses.
- **[Red Hat OpenShift Service on AWS (ROSA)](https://blowstack.com/tools/cloud-services-hub/red-hat-openshift-service-on-aws-(rosa))**: Amazon Managed Service for Prometheus can receive metrics forwarded from ROSA's Prometheus monitoring stack through authenticated remote write. This provides managed storage and querying for metrics collected from OpenShift workloads.

## Amazon Managed Service for Prometheus exam questions

Real practice exam questions on Amazon Managed Service for Prometheus, with the correct answer and its explanation.

### Question 1 (MLA-C02 (beta))

**Question:**

An ML platform team runs hundreds of model-serving deployments on Amazon EKS. A Prometheus-compatible managed metrics workspace already receives request counters from every pod at 15-second intervals. Several Amazon Managed Grafana dashboards refresh every minute and independently execute the same expensive PromQL calculation: the per-model request rate over the preceding 5 minutes, summed across pods. Dashboard queries become slow during peak traffic.

The team must reduce repeated query computation while keeping the displayed rates no more than 1 minute behind incoming metrics. Raw metrics must remain available at their existing resolution for troubleshooting. The team must not operate additional monitoring servers.

Which solution best meets these requirements?

**Correct answer:** Create a recording rule in the Amazon Managed Service for Prometheus workspace that evaluates the per-model request-rate expression every minute. Update the Grafana panels to query the recorded time series.

**Explanation:** Recording rules periodically evaluate PromQL expressions and store their results as new time series. Evaluating the shared aggregation once per minute lets dashboards reuse the precomputed rates instead of repeatedly processing the raw pod-level counters. The original 15-second metrics remain available, and the managed workspace runs the rules without additional servers.

### Question 2 (SAP-C03)

**Question:**

A company runs microservices on three Amazon EKS clusters in one AWS Region. Each cluster has two self-managed Prometheus collectors for high availability. Both collectors independently discover and scrape every application metrics endpoint in their cluster at the same interval. The collectors must remain redundant; dividing scrape targets between them is not permitted.

The company wants to replace its self-managed metrics storage with a managed, Prometheus-compatible backend. Existing PromQL dashboards must query metrics from all three clusters through one endpoint. Redundant collection must not double-count application metrics, and ingestion must support failover when a collector becomes unavailable. Metrics from different EKS clusters must remain distinguishable. The company does not want to operate an intermediary deduplication service.

Which solution meets these requirements?

**Correct answer:** Create one Amazon Managed Service for Prometheus workspace. Configure both collectors in each EKS cluster to remote-write to the workspace with SigV4 authentication. Set the external label cluster to a value unique to each EKS cluster and shared by its collector pair. Set __replica__ to a distinct value for each collector in the pair.

**Explanation:** Amazon Managed Service for Prometheus supports high-availability deduplication using the cluster and __replica__ labels. The shared cluster value identifies an HA group, and __replica__ distinguishes its redundant collectors. A different cluster value for each EKS cluster prevents unrelated sources from being grouped together. The service ingests data from the elected replica and supports failover, while one workspace provides a common PromQL query endpoint.

### Question 3 (SOA-C03)

**Question:**

A company runs two self-managed Prometheus servers for high availability. Both servers scrape the same application targets in an Amazon EKS cluster and use remote_write to send metrics to the same Amazon Managed Service for Prometheus workspace. Each server can handle collection for all targets if the other server fails.

A CloudOps engineer must configure the servers so that the workspace does not double-count samples from the two collectors and continues ingesting metrics when either collector fails. Both collectors must remain active, and application metric definitions must remain unchanged.

Which external-label configuration meets these requirements?

**Correct answer:** Set cluster="production" on both servers. Set __replica__="collector-1" on the first server and __replica__="collector-2" on the second server.

**Explanation:** Amazon Managed Service for Prometheus uses the cluster label to identify a high-availability group and __replica__ to distinguish its collectors. A shared cluster value and unique replica values enable the workspace to deduplicate samples and switch to the other replica if the selected replica stops sending metrics.

### Question 4 (DOP-C02)

**Question:**

A company runs production applications in two Amazon EKS clusters. Each cluster has two self-managed Prometheus servers for high availability. Both servers in a cluster independently scrape the same complete set of targets. The collected metrics do not already contain labels named cluster or __replica__.

A DevOps engineer must move long-term metric storage to a fully managed Prometheus-compatible backend while retaining the existing scrapers. Both servers must continuously forward metrics so that collection can continue if either server fails. Queries must not count redundant scrapes twice, and metrics from the two EKS clusters must remain separate. Network connectivity and IAM permissions are already configured.

Which configuration meets these requirements?

**Correct answer:** Configure both servers in each EKS cluster to remote write to one Amazon Managed Service for Prometheus workspace. Use the same cluster external-label value for the two servers in an EKS cluster, a different cluster value for the other EKS cluster, and a unique __replica__ external-label value for each server.

**Explanation:** Amazon Managed Service for Prometheus uses the cluster and __replica__ labels to identify high-availability collection groups and their replicas. Sharing a cluster value within each redundant pair enables deduplication and replica failover. Using a different cluster value for the other EKS cluster keeps its independent collection group and metrics separate.

## Practice Amazon Managed Service for Prometheus in a full exam

Questions on Amazon Managed Service for Prometheus in each practice exam, with every answer option and an explanation for each:

- [MLA-C02 (beta)](https://blowstack.com/practice-exams/aws-certified-machine-learning-engineer-associate): 5 questions
- [SAP-C03](https://blowstack.com/practice-exams/aws-certified-solutions-architect-professional): 3 questions
- [SOA-C03](https://blowstack.com/practice-exams/aws-certified-cloudops-engineer-associate): 2 questions
- [DOP-C02](https://blowstack.com/practice-exams/aws-certified-devops-engineer-professional): 1 question

## What is Amazon Managed Service for Prometheus

Amazon Managed Service for Prometheus is a fully managed, Prometheus-compatible monitoring service designed to efficiently monitor containerized applications at scale. Built to offer a seamless experience for developers and DevOps engineers who are already familiar with the power and flexibility of Prometheus, this service is optimized for monitoring metrics generated not just from Amazon EC2 instances but also from container services such as Amazon Elastic Kubernetes Service (EKS) and Amazon Elastic Container Service (ECS).

Amazon Managed Service for Prometheus runs as a serverless monitoring service.

The Amazon Managed Service for Prometheus collector is an agentless scraper that automatically discovers and monitors Amazon EKS applications and infrastructure without requiring you to manage Prometheus collection agents.

Prometheus has been the de-facto open-source standard for metric collection and alerting in Kubernetes environments, and by offering a managed service, Amazon brings the reliability and scalability of AWS infrastructure to the Prometheus ecosystem. One of the most significant benefits of using Amazon Managed Service for Prometheus is the reduction of operational complexity. Traditionally, maintaining a Prometheus setup involves configuring the Prometheus server, setting up storage, scaling the infrastructure to accommodate an increasing amount of metrics, and managing data retention policies. This service abstracts these complexities, allowing users to focus on defining metrics, setting up alerts, and gaining insights from their application's performance without worrying about the underlying infrastructure.

Workspace metrics are retained for 150 days by default and then automatically deleted. You can configure retention up to 1095 days (three years).

Moreover, Amazon Managed Service for Prometheus is designed to be scalable and secure. It can automatically scale its storage and query processing capabilities to meet the demands of thousands of container instances producing millions of metrics. This scalability is critical for modern applications that can experience variable load and need the monitoring system to adapt quickly. In terms of security, the service integrates with AWS Identity and Access Management (IAM), enabling fine-grained access control policies for different users and applications accessing the metrics data.

Active-series capacity per workspace automatically adjusts from 2 million to 50 million based on the last 30 minutes of usage. The active-series quota is adjustable up to 1.5 billion; a series counts as active when a sample has been reported within the past two hours.

Another notable feature is its compatibility with the Prometheus Query Language (PromQL), which offers users the flexibility to query their metrics in a highly expressive manner without needing to modify their existing dashboards and alerts. This compatibility ensures that teams can migrate to Amazon Managed Service for Prometheus without significant changes to their monitoring configurations.

The service exposes two APIs: an AWS API for managing resources such as workspaces, rule groups, and alert managers, and a Prometheus-compatible API for working within a workspace.

The integration with other AWS services further enhances the value of this managed service. For instance, it can be used alongside Amazon Managed Grafana, a complementary service that provides a powerful visualization tool for metrics, logs, and traces. This integration allows users to create comprehensive dashboards that aggregate and display data from multiple sources, offering a unified view of an application's performance and health.

In summary, Amazon Managed Service for Prometheus brings together the scalability, reliability, and security of AWS with the flexibility and popularity of Prometheus. It addresses the operational challenges of running a Prometheus server at scale while preserving the core functionalities and open-source ethos that have made Prometheus a favorite among developers and operators of containerized applications. Whether you're running a small setup or a large-scale Kubernetes cluster, this service provides a robust solution for monitoring your applications' performance and ensuring they run smoothly.

## Learn Amazon Managed Service for Prometheus in lessons

9 short lessons on Amazon Managed Service for Prometheus in 4 certification courses.

- MLA-C02 (beta): Amazon Managed Service for Prometheus: querying metrics for dashboards (Premium), https://app.blowstack.com/lessons/amazon-managed-service-for-prometheus-querying-metrics-for-dashboards-mla-c02
- SAP-C03: Amazon Managed Grafana vs Amazon Managed Service for Prometheus (Premium), https://app.blowstack.com/lessons/amazon-managed-grafana-vs-amazon-managed-service-for-prometheus-sap-c03
- SOA-C03: Amazon Managed Service for Prometheus: alerting rules (Premium), https://app.blowstack.com/lessons/amazon-managed-service-for-prometheus-alerting-rules-soa-c03
- DOP-C02: Amazon Managed Service for Prometheus: workspace metric ingestion (Premium), https://app.blowstack.com/lessons/amazon-managed-service-for-prometheus-workspace-metric-ingestion-dop-c02
- MLA-C02 (beta): Amazon Managed Service for Prometheus vs Amazon CloudWatch: metrics backends (Premium), https://app.blowstack.com/lessons/amazon-managed-service-for-prometheus-vs-amazon-cloudwatch-metrics-backends-mla-c02
- SAP-C03: Amazon Managed Service for Prometheus: resilient container metric ingestion (Premium), https://app.blowstack.com/lessons/amazon-managed-service-for-prometheus-resilient-container-metric-ingestion-sap-c03

## Concepts related to Amazon Managed Service for Prometheus

- [Metric Cardinality](https://blowstack.com/tools/cloud-concepts-hub/metric-cardinality): Metric cardinality is the number of distinct time series created by metric names and their unique combinations of dimension or label values.
- [Temporal Aggregation](https://blowstack.com/tools/cloud-concepts-hub/temporal-aggregation): Temporal aggregation combines timestamped observations within defined time intervals into summary values such as counts, sums, averages, or percentiles.
- [Time Series](https://blowstack.com/tools/cloud-concepts-hub/time-series): A time series is a sequence of observations indexed by time that represents how a measured quantity changes.
