Caddy

Monitor HTTP requests, errors, traffic, and resource usage from your local Caddy server. The agent probes the Caddy admin API to collect real-time performance data, helping you track traffic spikes and server health.

Setup

Preparing Caddy for monitoring requires ensuring the admin API is enabled and accessible to the agent.

1. Enable the admin API: By default, Caddy enables the admin API on localhost:2019. If you have disabled it or changed the port, ensure it is active.
2. Global options: You can explicitly enable metrics in your Caddyfile using global options:

   {
       admin 127.0.0.1:2019
   }

3. Restart Caddy: Apply any configuration changes by restarting or reloading Caddy.
4. Restart the agent: Restart simob to trigger service discovery: sudo systemctl restart simob.

Configuration

The agent automatically discovers Caddy metrics if they are exposed on the default http://localhost:2019/metrics endpoint. No additional configuration is required.

Metrics

The following metrics are collected from the Caddy admin API:

Metric	Type	Unit	Description
caddy_http_requests_total	Gauge	reqs	Total number of HTTP requests processed.
caddy_http_requests_rate	Gauge	reqs/s	Rate of HTTP requests per second.
caddy_http_errors_total	Gauge	errs	Total number of HTTP errors (4xx and 5xx).
caddy_http_errors_rate	Gauge	errs/s	Rate of HTTP errors per second.
caddy_http_response_bytes	Gauge	bytes	Total bytes sent in HTTP responses.
caddy_http_response_bps	Gauge	B/s	Throughput of response data in bytes per second.
caddy_http_request_duration_ms	Gauge	ms	Total time spent processing requests.

Troubleshooting

Verify the endpoint

Test locally using curl to ensure Caddy is correctly exposing statistics:

curl http://localhost:2019/metrics

Discovery check

If Caddy metrics are not appearing, verify that the admin API is bound to 127.0.0.1 and that the agent has permission to access it.