NGiNX provides various metrics thru their status page (requires the stub status module). Setting up monitoring for NGiNX with Nixstats is easy, you can read more at our helpdesk.
The most important metrics to watch and alert to for NGiNX are
requests_per_second. Besides this we recommend setting up alert through logging. You might want to setup a daily alert which will alert you of any new entries in the error logs of nginx.
The following metrics are available from the NGiNX monitoring plugin. Source code is available on GitHub.
The amount of accepted client connections per second.
The amount of handled client connections per second.
Amount of requests per second processed.
Total amount of accepted client connections.
Total amount of handled client connections.
Total amount of requests processed.
Current number of connections where nginx is reading the request header.
Current number of connections where nginx is returning the response back to the client.
Current number of idle client connections.
Current number of active client connections.
The top requested paths. Requires logging to be setup.
HTTP status codes parsed from your access logs. Requires logging to be setup.
NGiNX error logs entries. Requires logging to be setup.