Skip to main content
POST
Sends a JSON-RPC ping to every server with status = active and reports back which ones answered. Probes run in parallel via curl_multi_init (with a sequential wp_remote_post fallback when curl multi isn’t available), so latency is dominated by the slowest server, not the sum.
This is distinct from the public /health endpoint/health reports on the plugin itself and is unauthenticated. This endpoint is admin-only and reaches out to each MCP server to confirm the JSON-RPC transport is responsive.
Each per-server probe has a 5-second timeout and a 3-second connect timeout. Servers that exceed these limits are reported with status: "error" and the curl error string.

Body

No body. Send an empty POST.

Response Fields

array
One entry per active server. Empty when no servers are active.
string
Server name (the human-readable label, not the slug or UUID).
string
healthy when the server returned an HTTP status < 400. error for any non-2xx/3xx status or transport failure.
integer
HTTP status returned by the server. Present on every probe that completed at the HTTP layer (even errors like 502).
string
Transport-level error message — only present when the probe failed before getting an HTTP response (DNS, connection refused, timeout, TLS).