bus-api-provider-node

Node API Provider

bus-api-provider-node is the running-Bus HTTP surface for node capability discovery. Node mutation work is handled through bus-integration-node and remote command execution is delegated to the SSH runner integration.

Run this provider on loopback or a protected internal network behind bus-api. The current provider handler is an internal component and does not perform public-client authentication by itself; production deployments must mount this provider under bus-api internal routes matching /api/internal/node/*, require an operator/internal token with node administration scope at the Bus API layer, and keep the provider listener on loopback or a private service network. The deployment route should forward /api/internal/node/* to http://127.0.0.1:8093, and the Bus API auth policy should require an operator token with node:admin or the deployment’s equivalent internal node administration scope. Automation obtains its internal caller context from the deployed Bus auth/token flow or from bootstrap-only internal credentials managed outside Git.

Available endpoints return 200 with JSON on success:

  • GET /healthz returns {"ok":true,"service":"bus-api-provider-node"}.
  • GET /api/internal/node/capabilities returns generic node capability metadata with transport, events, and direct-bootstrap support.
bus-api-provider-node --addr 127.0.0.1:8093

In a second shell, verify the provider:

curl -fsS http://127.0.0.1:8093/healthz
curl -fsS http://127.0.0.1:8093/api/internal/node/capabilities

The first check succeeds with ok: true; the second includes node event names such as bus.node.bootstrap.request. Unknown paths return the normal HTTP 404 response from the provider mux.

For protected access through bus-api, enable the built-in provider by name and enable the matching module mount. Store the Bus API capability token in an untracked operator secret file before starting the service, then validate it is non-empty. This starts the node provider in the bus-api process; it does not require a separate bus-api-provider-node --addr 127.0.0.1:8093 process. The standalone provider command above is only for loopback diagnostics.

cd /path/to/deployment-repository
install -m 700 -d ./local
git check-ignore -q ./local/bus-api-capability-token || printf '%s\n' '/local/' >> .git/info/exclude
git check-ignore -q ./local/bus-api-capability-token
test -s ./local/bus-api-capability-token || openssl rand -hex 32 > ./local/bus-api-capability-token
chmod 600 ./local/bus-api-capability-token
test -s ./local/bus-api-capability-token
BUS_API_CAPABILITY_TOKEN="$(tr -d '\r\n' < ./local/bus-api-capability-token)"
test -n "$BUS_API_CAPABILITY_TOKEN"
bus-api serve --token "$BUS_API_CAPABILITY_TOKEN" --port 8080 \
  --provider node \
  --enable-module node

Keep bus-api serve running in that shell. In a second shell, set $BUS_API_BASE_URL to the local capability URL printed by bus-api, then verify the protected route:

cd /path/to/deployment-repository
BUS_API_CAPABILITY_TOKEN="$(tr -d '\r\n' < ./local/bus-api-capability-token)"
test -n "$BUS_API_CAPABILITY_TOKEN"
export BUS_API_BASE_URL="http://127.0.0.1:8080/${BUS_API_CAPABILITY_TOKEN}/v1"
curl -fsS "$BUS_API_BASE_URL/api/internal/node/capabilities"

With the correct Bus API capability URL, the command returns capability JSON. Without that URL token, the Bus API layer returns a 404 before it reaches this provider.

Use this provider on internal Bus routes. End users do not call it directly; deployment and operator automation call it with the appropriate internal authorization.