canonical: https://jentic.com/apis/amazonaws.com/amp # AWS Amazon Prometheus Service Jentic publishes the only available OpenAPI specification for Amazon Managed Service for Prometheus, keeping it validated and agent-ready. Amazon Managed Service for Prometheus (AMP) is a fully managed Prometheus-compatible monitoring service for container and Kubernetes workloads. The control-plane API exposes 21 operations covering workspaces, alert manager definitions, rule group namespaces, logging configuration, and resource tagging. Platform and SRE teams use it to provision metric stores at scale without operating their own Prometheus servers. ## For AI agents Provision and manage Prometheus-compatible metric workspaces, alert manager definitions, and rule groups for cloud-native applications via 21 control-plane operations. ## Scope Does not handle metric ingestion via remote-write, PromQL queries, or dashboard rendering - use for control-plane management of Prometheus workspaces, alert manager definitions, and rule groups only. ## Capabilities - Create and tag Prometheus workspaces that ingest remote-write metrics from EKS, ECS, or self-managed Kubernetes clusters - Define and update alert manager configurations to route firing alerts to receivers like Slack and PagerDuty - Manage rule group namespaces that hold recording and alerting rules evaluated server side - Toggle logging on a workspace to capture rules engine evaluation diagnostics in CloudWatch Logs - List and inspect workspaces with their status, tags, and creation timestamps for cost and capacity tracking ## Use cases ### Multi-Tenant Kubernetes Monitoring Platform teams use AMP to provide isolated Prometheus workspaces per tenant or environment without operating Prometheus servers. CreateWorkspace returns a remote-write endpoint that each cluster ships metrics to via Prometheus or the AWS Distro for OpenTelemetry. Tenants get isolated metric storage with shared infrastructure costs, and the workspace tag set is used for chargeback. Example prompt: Call CreateWorkspace with alias=tenant-acme and tags including CostCenter, then return the prometheusEndpoint for the cluster's remote-write configuration ### Centralised Alert Manager Configuration SRE teams use PutAlertManagerDefinition to push a single alert manager YAML configuration to many workspaces, ensuring consistent routing of firing alerts to on-call channels. The definition is versioned per workspace, and DescribeAlertManagerDefinition lets a deploy pipeline diff the active definition against the desired state before applying changes. Example prompt: Call PutAlertManagerDefinition with the YAML config Base64 encoded for the target workspace, then DescribeAlertManagerDefinition to confirm the new revision is active ### Recording and Alerting Rule Lifecycle Application teams ship recording rules and alerting rules to AMP through CreateRuleGroupsNamespace and UpdateRuleGroupsNamespace. Each namespace holds a set of rule groups defined in Prometheus YAML, and changes are evaluated server side as soon as they are applied. ListRuleGroupsNamespaces gives an inventory across a workspace for review and cleanup. Example prompt: Call CreateRuleGroupsNamespace with name=app-rules and a Base64 encoded YAML body containing recording and alerting rules for the workspace ### AI Agent Observability Provisioner An AI agent invoked through Jentic provisions a fresh Prometheus workspace whenever a new tenant onboards. The agent searches Jentic for the create-workspace operation, executes it with the tenant's tags, and returns the resulting endpoint URL to the deployment pipeline. Jentic handles SigV4 signing so the agent never sees the raw access key id and secret. Example prompt: Search Jentic for create amazon prometheus workspace, load the CreateWorkspace operation, execute it with the tenant alias and tags, and return the prometheusEndpoint ## Key endpoints | Method | Path | Description | | --- | --- | --- | | POST | /workspaces | Create a new Prometheus workspace | | GET | /workspaces | List all Prometheus workspaces in the region | | GET | /workspaces/{workspaceId} | Describe a workspace by id | | PUT | /workspaces/{workspaceId}/alertmanager/definition | Put an alert manager definition for a workspace | | POST | /workspaces/{workspaceId}/rulegroupsnamespaces | Create a rule group namespace inside a workspace | | PUT | /workspaces/{workspaceId}/logging | Enable or update logging for a workspace | ## Key resources - **Workspaces** — Create, describe, list, tag, and delete Prometheus workspaces that store ingested metrics - **Alert Manager Definitions** — Put, get, and delete alert manager YAML configurations per workspace - **Rule Group Namespaces** — Manage namespaces of recording and alerting rules evaluated against workspace metrics - **Logging Configuration** — Enable, update, and describe rules-engine logging for a workspace - **Tags** — Tag workspaces and other resources for cost allocation and ownership tracking ## Why Jentic - **Setup:** Wiring Amazon Managed Service for Prometheus by hand means signing every control-plane request with AWS Signature Version 4 and resolving the aps.{region}.amazonaws.com host for your region. Through Jentic you install once, import Amazon Prometheus Service from the API Directory, store your AWS access key and secret once, and your agent calls it. - **Permission scoping:** AMP puts the workspace id in the URL path (/workspaces/{workspaceId}/...), so a rule can pin the agent to one Prometheus workspace and its alert-manager and rule-group configuration. You choose the operations it may call, so actions like creating new workspaces with POST /workspaces are not included unless you add them. - **Credential handling:** Your AWS access key and secret are stored once, encrypted, by your own Jentic One instance and injected at execution time. They never enter the agent's prompt, logs, or context. - **Discovery method:** Agents search Jentic by intent such as 'create a Prometheus workspace' or 'update the alert manager definition', and Jentic returns the matching AMP operation with its input schema so the agent does not have to map operation names to AWS path conventions. ## Related APIs - **Amazon API Gateway** — API Gateway emits metrics that can be relayed into AMP for unified observability of public APIs - **AWS Amplify** — Amplify-hosted apps can ship custom metrics into an AMP workspace for production monitoring - **New Relic** — Full observability platform with Prometheus remote-write support; choose when you also need APM and logs in one tool ## FAQ ### Why is there no official OpenAPI spec for Amazon Managed Service for Prometheus? AWS does not publish an OpenAPI specification for AMP; the official surface is the AWS SDKs and the Smithy model. Jentic generates and maintains this spec so that AI agents and developers can call Amazon Prometheus Service via structured tooling. It is validated against the live API and kept up to date. Get started with Jentic One, the self-hosted execution layer. ### What authentication does the Amazon Managed Prometheus API use? AMP uses AWS Signature Version 4 (SigV4) on the control plane, signed with an IAM access key id and secret access key. The data plane (remote-write and query) uses SigV4 too. Through Jentic the access key id and secret access key are stored encrypted in the vault and SigV4 signing is performed server side. ### Can I push Prometheus rules with the Amazon Managed Prometheus API? Yes. Use POST /workspaces/{workspaceId}/rulegroupsnamespaces with a Base64 encoded YAML body containing one or more rule groups. AMP evaluates the rules server side as soon as the namespace is created, and PUT updates an existing namespace in place. ### What are the rate limits for the Amazon Managed Prometheus API? AMP applies per-region throttling on control-plane operations like CreateWorkspace and PutAlertManagerDefinition; rates are higher for read operations like ListWorkspaces. Data-plane remote-write and query throttling depend on the workspace's active series and query rate. ThrottlingException should be retried with exponential backoff. ### How do I create a workspace with the Amazon Managed Prometheus API through Jentic? Install the SDK with pip install jentic, search Jentic for create amazon prometheus workspace, load the CreateWorkspace operation schema, and execute it with an alias and tags. Jentic handles SigV4 signing and returns the workspace id and remote-write endpoint URL. ### Can I delete a workspace and recover its metrics later? No. DeleteWorkspace permanently removes the workspace and all stored metrics; there is no undelete. Before deletion, snapshot any required dashboards or export metrics through the query API. AMP retains metrics for the workspace's configured retention period only while the workspace exists. ### Can I limit what my agent is allowed to do with the Amazon Managed Prometheus API? Yes. Jentic One runs self-hosted, so you set the rules that decide which AMP operations and credentials your agent may use. Because AMP puts the workspace id in the URL path (/workspaces/{workspaceId}/...), you can pin the agent to a single Prometheus workspace and its alert manager and rule group configuration. You pick the operations it may call, so broad actions like creating new workspaces with POST /workspaces stay off limits unless you add them.