Choosing a channel
Four distinct channels are available and each suits a different purpose. Using
the wrong one is the most common source of integration friction.
| Need | Channel |
|---|---|
| Alarm on node up/down and capacity | GET /api/bng/health, polled every 5–15 s |
| Standard metric collection into an existing TSDB | Scrape /metrics with a scoped key |
| Charts and historical queries | /throughput/recent, /metrics/query_range, or /analytics/v1/* |
| Per-subscriber traffic and QoS detail | The /v1/* passthrough — raw telemetry |
| Fault console with acknowledgement | /api/bng/events plus the DELETE acknowledgement |
| "Who changed what, when" | /api/bng/audit/timeline |
| Continuous compliance feed into a SIEM | Configure an audit shipping destination |
| Change control and rollback | /api/bng/snapshots with dry-run restore |
The general principle: for anything about subscribers, use the raw telemetry
plane; for anything about the node's operation, use this one.