Dwara

Appearance

Observability and analytics ​

Two complementary views onto your traffic. Observability answers "is it healthy right now" - the live signals (logs, metrics, request IDs, error envelopes) an operator uses during an incident. Analytics answers "what happened over time, to whom, and why" - the durable, queryable history used after the fact. The two are independent: run either, both, or neither.

Alert webhooks sit alongside these as the push channel for gateway state changes (breaker trips, endpoint ejections, config publishes).

In this section ​

  • Observability - structured logs, request IDs, the metrics families, SLOs and error budgets, and the JSON error envelope.
  • Analytics - the embedded analytics store: durable traffic history with rollups, retention, and bounded disk usage.
  • Analytics stream - the opt-in raw firehose out to an external HTTP collector, fire-and-forget end to end.
  • Alert and event webhooks - POST JSON notifications on gateway state changes to your incident tool or collector.

Runnable demo ​

The demos/06-observability/ directory in the repository runs a live stack for the features in this section; its README covers prerequisites, test scripts, and teardown.