Skip to main content
Glama

Connect metrics source

connect_metrics_source
Destructive

Propose connecting the Prometheus-compatible metrics source Ankra derived from this cluster's installed addons (the one a query_prometheus refusal named). Pass EXACTLY the provider and endpoint from that refusal - both are re-derived server-side at execution and the call is refused on any mismatch, so never invent or edit them. The action requires an organisation admin's confirmation; on confirm the endpoint is first verified with the existing read-only probe (a failed probe writes nothing), then the source is registered in the resource graph as the cluster's prometheus_metrics source, inheriting GitOps, audit, drift detection and IaC export. An in-cluster Service endpoint needs no credential - the agent relay dials it from inside the cluster - so never ask the user for one. Do not use this to repair a misconfigured existing source. If the store runs OUTSIDE the cluster, or the derived endpoint does not match what the release actually renders, re-propose the correct endpoint with endpoint_is_operator_supplied=true: derivation only sees installed addons, so neither case can be derived and neither is a reason to refuse the connection. Admin confirmation and the read-only probe apply either way.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "provider": "<provider>",
      "endpoint": "<endpoint>",
      "cluster_id": "<cluster_id>"
    },
    "name": "connect_metrics_source"
  }
}

Input Schema

NameRequiredDescriptionDefault
endpointYesThe endpoint URL. Copy verbatim from the refusal's derived-candidate sentence, or supply the correct one with endpoint_is_operator_supplied=true when the store is external or the derived endpoint is wrong. Must be an http:// or https:// URL. Private and in-cluster addresses are fine; loopback, link-local (including the cloud metadata address) and multicast are refused.
providerYesThe provider. Copy verbatim from the refusal's derived-candidate sentence, or supply it with endpoint_is_operator_supplied=true.
cluster_idYesAnkra cluster id (UUID) the call targets. Discover ids with list_clusters.
endpoint_is_operator_suppliedNoSet true ONLY when a person supplied this endpoint rather than copying Ankra's derived candidate. Derivation reads INSTALLED ADDONS, so it cannot name a store running outside the cluster (a managed OpenSearch, a hosted Prometheus) - that is the case this exists for. It skips the derived-candidate equality check ONLY: organisation-admin confirmation and the read-only probe still run, and the confirmation shows the endpoint as supplied rather than derived. Never set it to pass off an endpoint you inferred yourself.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.8/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only declare destructiveHint=true and idempotentHint=false; the description adds the whole behavioral picture: server-side re-derivation with refusal on mismatch, org-admin confirmation, a read-only probe that writes nothing on failure, registration into the resource graph with GitOps/audit/drift/IaC inheritance, and the fact that no credential is needed for an in-cluster Service endpoint. That is substantial context the annotations cannot carry.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the action and the copy-verbatim constraint, then ordered into confirmation, probe, registration and exceptions. It is dense and long, and the closing line about admin confirmation and the probe applying either way partly restates what the schema's endpoint_is_operator_supplied description already says.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a destructive, confirmation-gated registration with no output schema, the description covers preconditions, failure semantics, credential expectations, and the two edge cases that would otherwise cause a wrong refusal. Nothing an agent needs in order to call it correctly is missing.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so the baseline is 3, but the description adds a real invariant the schema does not state: provider and endpoint are re-derived server-side at execution and the call is refused on any mismatch, so they must be copied verbatim. It also clarifies that operator-supplied endpoints skip only the equality check, not confirmation or the probe.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb (propose connecting) and resource (the Prometheus-compatible metrics source Ankra derived from installed addons), plus the exact provenance of the inputs (the one a query_prometheus refusal named). An agent can distinguish it from connect_log_source and from repairing an existing source without opening a schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives explicit when-to-use, explicit when-NOT-to-use ("Do not use this to repair a misconfigured existing source"), and routes two named exceptions (external store, derived endpoint mismatch) to the operator-supplied path rather than to refusal. Alternatives and exclusions are both present.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.