Claude Control Plane connectorWhat Claude can do in your Control Plane account.

Browses the catalog of ready-to-install templates: name, category, latest version, and whether each one brings its own GVC.

Translates a resource manifest into Terraform code, after a dry-run check against the API.

Generates Terraform code for resources already in place, one at a time or in bulk, from their link.

Fetches one asynchronous command by its ID, to see where a long-running operation stands: in progress, done, or failed, without relaunching anything.

Loads the platform's operating guide: its resource model, production defaults, and how to verify that an operation actually landed.

Loads the runbook for one family of tasks: how to use a feature properly, the constraints that are easy to miss, and when it's the wrong tool.

Reads a build's status, progress events, and log from its build ID.

Shows an installed template's status, its revision number, and the full set of resources tied to it on the platform.

Lists the permissions that can be granted on a resource kind, to build an accurate access policy.

Fetches one resource by kind and name, with its full definition in JSON, so Claude sees every setting exactly as the platform stores it.

Returns the exact schema of a resource kind and its API endpoints, for drafting a correct manifest.

Shows a catalog template's available versions, its prerequisites, and an example values file to start from.

Fetches one trace by ID and summarizes its span tree, durations, and errors.

Fetches a workload's event log to diagnose a failed start, a failing health check, or a stuck deployment.

Queries a workload's logs through simple parameters or a raw LogQL query.

Lists the asynchronous commands issued against a workload or a volume set: cron runs, replica stops, volume operations. Claude finds the ID to follow up on there.

Reports a workload's rollout status location by location. It's the first check after going live.

Lists the template releases present in an organization, so Claude can tell what came from the catalog and what was set up another way.

Catalogs the metrics and labels you can query, each with a correct PromQL template, the query language for metrics. Claude starts there so it doesn't aim at a metric that doesn't exist.

Lists the organization's resource quotas and current usage, with a filter for those close to the limit.

Builds a summary table of the resources of one kind, workloads, domains, or volumes, to give Claude the big picture before digging in.

Lists a workload's running replicas in one location, meaning the copies of the app actually serving traffic right now.

Queries the audit trail of operations performed on a resource kind, with who did what and when.

Runs a PromQL query against the platform's metrics and returns the values, which Claude can then comment on, compare, or sum up in one sentence.

Searches distributed traces, in TraceQL, for slow or failing requests. A trace follows a request's full path through your services.

Searches Control Plane's documentation base: guides, examples, and API references. Claude leans on it so it doesn't answer about the platform from memory.