Prometheus Metrics

GET /metrics    (on metrics port :9093, NOT the API port)

Note: Metrics are served on a separate, dedicated port (:9093 by default, configurable via OSPREY_METRICS_LISTEN). This keeps operational telemetry isolated from the authenticated API surface on :8080.

Available metrics (API server):

Metric	Type	Labels	Description
`osprey_http_requests_total`	counter	method, route, status	HTTP request count by method, route pattern, HTTP status
`osprey_http_request_duration_seconds`	histogram	method, route	HTTP request latency (buckets: 5ms–5s)
`osprey_db_pool_total_conns`	gauge	—	Total connections in the pgx pool
`osprey_db_pool_idle_conns`	gauge	—	Idle connections
`osprey_db_pool_acquired_conns`	gauge	—	Currently acquired connections
`osprey_db_pool_max_conns`	gauge	—	Maximum pool connections
`osprey_nats_messages_published_total`	gauge	—	Total NATS messages published
`osprey_nats_messages_received_total`	gauge	—	Total NATS messages received
`osprey_nats_reconnects_total`	gauge	—	Total NATS reconnections
`osprey_nats_connected`	gauge	—	NATS connection status (1/0)
`osprey_ws_clients_connected`	gauge	—	Connected WebSocket clients
`osprey_ws_messages_broadcast_total`	counter	—	Broadcast messages sent
`osprey_ws_messages_dropped_total`	counter	—	Messages dropped (slow clients)

Engine metrics (port 9090):

Metric	Type	Labels	Description
`osprey_engine_snapshots_total`	counter	area	Topology snapshots processed
`osprey_engine_snapshot_duration_seconds`	histogram	area	Snapshot processing time
`osprey_engine_snapshot_errors_total`	counter	—	Snapshot processing errors
`osprey_engine_devices_processed_total`	counter	—	Total devices processed
`osprey_engine_links_processed_total`	counter	—	Total links processed
`osprey_engine_stale_sweeps_total`	counter	—	Stale area sweep operations

Collector-manager metrics (port 9091):

Metric	Type	Labels	Description
`osprey_manager_collectors_running`	gauge	—	Running collector instances
`osprey_manager_collectors_errored`	gauge	—	Errored collector instances
`osprey_manager_sync_total`	counter	—	DB poll sync cycles
`osprey_manager_sync_errors_total`	counter	—	Failed sync cycles

SNMP discovery metrics (collector-manager port 9091):

Metric	Type	Labels	Description
`osprey_snmp_isis_area_verification`	gauge	collector, area, state	One-hot latest IS-IS area verdict (`unknown`, `verified`, `pending`, `unreachable`, `truncated`)
`osprey_snmp_isis_bootstrap_areas`	gauge	collector, state	Active handoff counts (`desired`, `acked`); both return to zero after handoff
`osprey_snmp_lsdb_walk_truncated_total`	counter	target, area, protocol	LSDB walks rejected after transport or structural completeness validation

Engine and collector-manager metrics also include the osprey_db_pool_* and osprey_nats_* gauges.