Run a PromQL range query through an Operator agent
Run a PromQL range query against a Prometheus reached through the Operator agent enrolled for clusterArn. Use this instead of the Amazon Managed Prometheus endpoint (aws/amp/query-range) when the module's ui.metrics prometheus source carries cluster_arn + endpoint — an in-cluster Prometheus, or an external one such as Grafana Cloud Metrics. Applies the same Redis answer cache the in-cluster Loki queries use. A cluster whose agent is disconnected answers 200 with no series and the ARN in offlineClusters.
Request
- Base URLhttps://api.ravion.com
- AuthHTTP bearer
Query parameters
ARN of the EKS cluster whose Operator agent proxies the query. Resolved to a connected agent inside the caller's own organization, which is what confines this endpoint to clusters the caller can see — an ARN belonging to another tenant resolves to nothing.
Base URL of the Prometheus to query, without the /api/v1/query_range suffix. Naming it is not a permission: the agent judges every proxied destination against its own httpProxy.allowedEndpoints, so an endpoint that install was not configured for is refused in the cluster.
Name of a Kubernetes Secret in the Operator namespace whose contents the AGENT presents as the Authorization header. A name, never a value — Ravion neither reads nor stores the credential. Omit for an endpoint that needs none.
PromQL expression. Occurrences of $__interval are replaced server-side with the normalized step, so range-vector windows always match the resolution actually queried.
Requested step in seconds. Widened server-side to a 15s floor and to whatever keeps the window under the 11,000-samples-per-series cap; the effective value is echoed as stepSeconds.
Inclusive start of the query window (ISO 8601). Must be earlier than endTime.
Inclusive end of the query window (ISO 8601). Must be after startTime and not in the future.
Response
The request has succeeded.
Changes
No recorded changes to this operation across all 2 revisions of this API.