What? A simple server monitoring tool.

What is a modular, easily-extensible server monitoring tool which is still in the early stages of development.

$ gem install what

Right now, the example config file monitors Unicorn workers. Observe:

$ what -c example/what.yml >/dev/null 2>&1 &
[1] 2392
$ curl localhost:9428
{"unicorn":{"details":[{"cpu_time":"0:00.02","pid":"11023"},{"cpu_time":"0:00.02","pid":"11022"}],"health":"ok","workers":2},"health":"ok"}
$ sudo /etc/init.d/unicorn stop
$ curl localhost:9428
{"unicorn":{"details":[],"health":"alert","workers":0},"health":"alert"}

When the health value of any module is set to "alert" instead of "ok", the HTTP request returns 503 instead of 200. This means What can easily be used in conjunction with monitoring tools like Pingdom.

Writing monitoring modules is easy: the only requirement is that they implement a health method, which returns 'ok', 'warning', or 'alert'. They can also implement a details method, which returns the hash that's included in the HTTP response. See the included What::Modules::Base and What::Modules::Unicorn classes for the implementation details.