Observability
Deep health, config validation, and an aggregate metrics summary on EdgeControl. The Prometheus /metrics endpoint sits on the SBI listener (and observability.metrics_listen when set), not on the admin gRPC port; see the Metrics reference.
Deep health
GET /admin/health/deep
Runs three coarse subsystem checks — whether any producers are configured, whether the control-plane store is reachable, and the state of the Diameter relay — and reports each result. On the CLI this is JSON from d3x-edgectl health-deep. For a Kubernetes readiness probe, hit GET /ready on the SBI or metrics listener instead of the admin gRPC port.
status is healthy, degraded, or unhealthy. Each entry under checks is ok, degraded, fail, or — for an unconfigured Diameter relay — disabled.
Config validation
POST /admin/config/validate
Validates a candidate config payload (YAML or JSON, selected by the Content-Type header) without applying it. Returns {"valid": true} on success, or {"valid": false, "errors": ["..."]} with one string per validation failure.
Metrics summary
GET /admin/metrics/summary
An aggregate snapshot of the registered metric families, rolled up for at-a-glance use. Distinct from the full /metrics Prometheus exposition on the SBI listener (see the Metrics reference).
Query parameters:
prefix— include only metric families whose name starts with this string.max_series— cap the number of emitted series (default 10000).
Series are grouped under families, keyed by metric name. When the max_series cap is hit, the response adds "truncated": true so a caller knows to narrow the request with prefix.
Examples
Where to go next
- Metrics reference — the full Prometheus series list.
- Observability guide — scrape, dashboards, alerts, tracing.
- Auditing and compliance — shipping structured decision logs.