Admin Tool Performance

Per-tool latency telemetry + the fleet performance KPI, so we can MEASURE and drive down tool slowness over time (not just cap it). Source: mcp_requests.latency_ms — the cloud-connector round-trip, i.e. the latency the user actually experiences. tools/call only.

KPI tool_latency_sla = % of tool calls that complete under slow_ms (default 10s). Higher is better; the goal is to push it up. The per-tool table (worst p95 first) says where to optimize.

get/admin/tool-performance

Request

  • Base URL: https://api.local-mcp.com (the document declares no server; this is the origin it was published from)
  • URL: https://api.local-mcp.com/admin/tool-performance
  • Auth: none declared. Required headers: x-admin-key

Query parameters

daysinteger
slow_msinteger

Headers

x-admin-keystring required

Response

Successful Response

unknown required

Changes

No recorded changes to this operation across all 1 revision of this API.