Amazon Managed Service for Prometheus

Amazon Managed Service for Prometheus is a fully managed, Prometheus-compatible monitoring and alerting service that makes it easy to monitor containerized applications and infrastructure at scale.

Features

• You can use the Prometheus query language (PromQL) to filter, aggregate, and alert on metrics for performance visibility.

• The service automatically scales as your monitoring needs grow.

• It is integrated with Amazon EKS, Amazon ECS, and AWS Distro for OpenTelemetry.

• It supports multi-AZ replication within an AWS Region for high availability.

• You can collect and query metrics from container clusters running on AWS and on-premises.

Concepts

Workspace

• A workspace is a logical and isolated space dedicated to storing and querying Prometheus metrics.

• You can have one or more workspaces in each AWS Region.

• Metrics ingested into a workspace are stored for 150 days.

• A workspace can support multiple rule files.

Rules

• To use rules in Prometheus, you create YAML-based rule files.

• Each rules file is contained within its own namespace.

• Rules are organized into rule groups within a file and are evaluated sequentially from top to bottom.

Recording Rules

• These rules allow you to precompute frequently used or computationally expensive expressions and save their results as a new set of time series.

• Querying the precomputed results is significantly faster than running the original complex expression every time.

Alerting Rules

• These rules allow you to define alert conditions using PromQL expressions and a threshold.

• When a rule's condition is met and the threshold is triggered, a notification is sent to the Alert Manager, which then forwards it to a configured receiver like Amazon SNS.

Alert Manager

• The Alert Manager handles alerts sent by Prometheus. It supports deduplication, grouping, and routing of alerts to the correct notification channel.

• You can also silence and inhibit alerts.

• Configuration File: The alert manager's behavior is defined in a configuration file which also stores templates.

• Grouping: Collects similar alerts into a single notification to reduce noise.

• Inhibition: Suppresses notifications for certain alerts if other specified alerts are already firing.

• Silences: Mute alerts for a specified period.

• Templating: Allows for the use of templates in alert notifications.

• The service supports tagging for both workspaces and rule group namespaces.

Pricing

• You are charged based on three metrics:

  1. Metrics Ingested: You are charged for the volume of metrics you send to Prometheus.

  2. Metrics Queried: PromQL Query Samples Processed (QSP) is charged on a per-billion sample basis.

  3. Metrics Stored: Storage fees are calculated based on the compressed size of metric samples and their metadata.

• There are no charges for inbound data transfer.