Module prometheus 

The prometheus node — a thin parity node over featherbit’s built-in Prometheus metrics.

featherbit already exposes Prometheus metrics out of the box. The GatewayMetrics registry records per-route request counters and latency histograms plus per-node execution metrics; the graph engine and data-plane listener feed it on every request with no plugin required, and it is rendered at the Admin API’s /metrics endpoint. So this node does not stand up metrics from scratch.

It exists only to add a dimension APISIX’s prometheus plugin tracks that featherbit’s always-on core metrics do not: a per-consumer request counter (gateway_consumer_requests_total, labelled consumer and route). Drop it into a pipeline after the auth node that attaches the consumer (key-auth, basic-auth, …); each execution bumps the counter for the request’s consumer (anonymous when none is attached). It never mutates the context and never fails.

Deviations from APISIX

• APISIX’s plugin wires up the entire Prometheus exporter and its full metric set. In featherbit the core metrics are built-in and always on, so this node is a thin add-on that only records the per-consumer counter.
• APISIX’s prefer_name toggles route name vs route id in labels. featherbit has no route object on the context here, so the route label uses the request Host (kept low-cardinality on purpose). prefer_name is accepted for config compatibility but is otherwise inert.

Structs

PrometheusPlugin

Records a per-consumer request counter against the shared GatewayMetrics; a no-op when metrics are disabled (unit tests).