Caddy Dashboard - Monitor Requests, Latency, Upstreams

This dashboard shows how Caddy handles traffic as a web server and reverse proxy. You can track request rate and status codes, request duration and time to first byte, handler errors, bandwidth, the health of reverse proxy upstreams, and config reloads.

The data comes from the metrics that Caddy v2.11.3 and newer push over OTLP when the metrics global option has otlp.

Dashboard Coverage

• Read the request rate and the share of 5xx responses.
• Break traffic down by route handler, status code, and HTTP method.
• Track request duration percentiles and time to first byte for each Caddy server.
• Find handler errors, such as an upstream that does not respond.
• Watch request and response bandwidth and payload sizes.
• Catch unhealthy reverse proxy upstreams and failed config reloads.

Metrics Included

Caddy records request metrics once for each route that a request matches, and labels them with the first handler of the route. To count each request once, the request panels read one route that sees every request, set by the route variable. Requests per Second by Handler is the exception: it shows every route so that you can pick one.

Overview

• Request Rate: caddy_http_requests_total
• 5xx Responses (%): caddy_http_request_duration_seconds.count, 5xx codes divided by all codes
• Requests In Flight: caddy_http_requests_in_flight
• Unhealthy Upstreams: caddy_reverse_proxy_upstreams_healthy, the number of upstreams that report 0

Traffic

• Requests per Second by Handler: caddy_http_requests_total by server and handler, for every route. Use it to find the route that sees every request.
• Requests per Second by Status Code: caddy_http_request_duration_seconds.count by code
• Requests per Second by Method: caddy_http_request_duration_seconds.count by method
• Requests In Flight by Server: caddy_http_requests_in_flight by server

Latency

• Request Duration (p50, p95, p99): caddy_http_request_duration_seconds
• p95 Request Duration by Server: caddy_http_request_duration_seconds by server
• p95 Request Duration by Status Code: caddy_http_request_duration_seconds by code
• p95 Time to First Byte by Server: caddy_http_response_duration_seconds by server

Errors

• 4xx and 5xx Responses per Second: caddy_http_request_duration_seconds.count by code
• Middleware Errors per Second by Server: caddy_http_request_errors_total by server

Bandwidth

• Request Bandwidth by Server: caddy_http_request_size_bytes.sum
• Response Bandwidth by Server: caddy_http_response_size_bytes.sum
• p95 Request Size by Server: caddy_http_request_size_bytes
• p95 Response Size by Server: caddy_http_response_size_bytes

Upstreams and Configuration

• Upstream Health: caddy_reverse_proxy_upstreams_healthy by upstream. Caddy reports 0 only when reverse_proxy has active or passive health checks.
• Last Config Reload Successful: caddy_config_last_reload_successful

Dashboard Variables

• route: The first handler of the route that sees every request. The default is tracing, which fits the setup guide. Without tracing, pick the handler of your first route that has no matcher. The request, latency, error, and bandwidth panels use it. Requests per Second by Handler does not.
• service_name: The service.name of your Caddy services, from OTEL_SERVICE_NAME. Every panel uses it.

Next Steps

• Set up Caddy to send traces, metrics, and access logs to SigNoz.
• Create metric-based alerts on the 5xx percentage or on unhealthy upstreams.
• Go Runtime Metrics covers the go_* metrics that Caddy also sends.