Skip to content

Health Monitoring

dawu-manager monitors the health of registered BNG nodes by querying the dawos-agent /health endpoint. Health check results determine the node's status indicator displayed throughout the dashboard.

How Health Checks Work

When a health check is triggered, dawu-manager:

  1. Decrypts the stored API key for the target node.
  2. Sends an HTTP GET request to <node-url>/health.
  3. Evaluates the response status and body.
  4. Updates the node's status and lastSeen fields in the database.

The health endpoint on dawos-agent returns:

{
  "status": "ok",
  "version": "0.3.3",
  "node_name": "bng-jakarta-dc1-01",
  "uptime": "5d 12h 34m"
}

Status Indicators

Status Color Condition
Online Green Node responded with HTTP 200 and status: "ok"
Offline Red Node is unreachable, timed out, or returned an HTTP error
Degraded Yellow Node responded but reported partial functionality issues
Unknown Gray Node has never been health-checked since registration

Manual Health Check

Click the Check Health button on any node card (dashboard) or node detail page to trigger an immediate health check. The status indicator updates in real time based on the response.

Health Check Data Flow

dawu-manager                              BNG Node
+-------------------+                     +-----------------+
| API Route         |  GET /health        | dawos-agent     |
| /api/nodes/[id]/  | -----------------> | :8470           |
|   health          |                     |                 |
|                   | <-- 200 OK -------- | {"status":"ok"} |
| Update DB:        |                     +-----------------+
|   status="online" |
|   lastSeen=now()  |
+-------------------+

Troubleshooting Unhealthy Nodes

When a node shows as Offline, investigate the following in order:

1. Verify dawos-agent Is Running

SSH into the BNG node and check the service status:

sudo systemctl status dawos-agent

If the service is stopped, start it:

sudo systemctl start dawos-agent

2. Test the Health Endpoint Locally

On the BNG node itself:

curl -sf http://localhost:8470/health

If this fails, the issue is with dawos-agent or its configuration, not with network connectivity.

3. Check Network Connectivity

From the dawu-manager host, test connectivity to the BNG node:

curl -sf http://<node-ip>:8470/health

If this times out, check:

  • Firewall rules on both the dawu-manager host and the BNG node (port 8470 must be open).
  • Network routing between the two hosts.
  • Any VPN or tunnel configurations that may affect connectivity.

4. Verify the API Key

Test authentication from the dawu-manager host:

curl -sf -H 'X-API-Key: <api-key>' http://<node-ip>:8470/api/v1/sessions

If this returns HTTP 401, the API key stored in dawu-manager does not match the key configured on the BNG node. Edit the node in dawu-manager and update the API key.

5. Check dawos-agent Logs

On the BNG node:

sudo journalctl -u dawos-agent --since '10 minutes ago' --no-pager

Look for startup errors, binding failures, or authentication rejection messages.

Health Check Timeout

Each health check request has a timeout of 10 seconds. If the dawos-agent instance does not respond within this window, the node is marked as Offline.

Common causes of timeouts:

  • High CPU load on the BNG node causing slow response.
  • Network congestion or packet loss between dawu-manager and the BNG node.
  • dawos-agent process is running but unresponsive (hung state).

Permissions

All authenticated users (viewer, operator, and admin) can view node health status and trigger manual health checks. No elevated role is required for read-only health monitoring.