Outputs

An output is a connected backend that receives the topology changes the engine produces. You declare outputs through the web UI or the REST API; the engine maintains the connection, publishes every change to every enabled output in parallel, and surfaces the connection state in real time.

What outputs do

Whenever the topology moves — a node appears, a link metric changes, a prefix is withdrawn — every enabled output receives a message describing the change. The message is the same structured JSON object you read from the web UI and the REST API, never a raw BGP-LS NLRI. See Topology / Change events for the semantics of what gets published.

Outputs run in parallel: a slow consumer never blocks fast ones, and adding or removing an output never disrupts existing ones.

Output types

NATS is the only output type in this release. Additional types (Kafka, webhook, file) are on the roadmap.

Managing outputs

From the web UI, open Outputs in the left navigation. The list shows each output with its name, type, connection URL, enabled state, and current connection state (disconnected, connecting, connected, error). Click an output to view its detail page, edit it, or delete it. Use New output to create a new one.

Outputs are also fully managed through the REST API under /api/outputs.

Connection state

Every output has a state the engine updates as it manages the underlying connection:

• disconnected — the output is disabled, or the engine has not yet attempted to connect.
• connecting — the engine is attempting to open the connection.
• connected — the connection is up and messages are flowing.
• error — the last connection attempt failed; last-error on the detail page carries the reason.

The state updates live in the UI and on the Outputs list.