Metrics
Four operational dashboards for the interoperability layer, read from GET /api/ops/metrics. That endpoint needs a staff token with the OFFICER or ADMIN role. No actuator or Prometheus endpoint is served.
Reading the metrics…
Connector health
Per data source: outcome of each source call (success, failure, unavailable) and exchange latency, retries included. Calls are counted since this node started; latency percentiles cover the last 10 minutes on this node. Latency is for successful exchanges unless the column says otherwise.
- Success
- Failure
- Unavailable (breaker open, bulkhead full, source killed, no adapter)
SLA compliance
Open applications (not approved, rejected or closed): SLA due time against now, from tracking. Read live from the database.
- Within SLA
- Due within 24 hours
- Breached
Consent and access
Every access-authorize decision, since this node started. A grant is counted when it is issued; a denial by its reason code. No citizen or consent identifiers are recorded.
Exception queue
Open orchestration exceptions: depth and age of the oldest (a gauge over the exception table, refreshed every few seconds). Retry them from Exceptions.