Monitoring¶
All pods expose metrics for Prometheus at the /metrics endpoint.
By default, the /metrics endpoint is hidden from public access via the
global.networkingConfiguration.hideMetrics setting (default: true).
This uses the NGINX Ingress Controller’s server-snippet annotation to restrict
access to internal networks only. If you use a different ingress controller,
you may need to set hideMetrics: false and handle access control separately.
To configure Prometheus scraping, point your Prometheus instance at the
individual service endpoints. The /stats endpoint (served by hub-stats)
provides a human-readable overview of the current system state, including
active connections per hub.