Simple HTTP-based system information server
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


Porkchop is a simple HTTP-based system information server. You write plugins for it and it responds with the data based on your request.

Here is an example:

scott@beatbox:~% curl http://localhost:5000/cpuinfo
/cpuinfo/processor2/fpu yes 
/cpuinfo/processor2/f00f_bug no
/cpuinfo/processor2/cache_alignment 64
/cpuinfo/processor2/vendor_id AuthenticAMD
/cpuinfo/processor2/flags fpu 
/cpuinfo/processor2/bogomips 6384
/cpuinfo/processor2/hlt_bug no
/cpuinfo/processor2/apicid 2
/cpuinfo/processor2/fpu_exception yes 
/cpuinfo/processor2/stepping 3
/cpuinfo/processor2/wp yes 
/cpuinfo/processor2/siblings 4
/cpuinfo/processor2/model 4
/cpuinfo/processor2/coma_bug no
/cpuinfo/processor2/fdiv_bug no
/cpuinfo/processor3/fpu yes 
/cpuinfo/processor3/f00f_bug no
/cpuinfo/processor3/cache_alignment 64
/cpuinfo/processor3/vendor_id AuthenticAMD
/cpuinfo/processor3/flags fpu 
/cpuinfo/processor3/bogomips 6384
/cpuinfo/processor3/hlt_bug no
/cpuinfo/processor3/apicid 3
/cpuinfo/processor3/fpu_exception yes 
/cpuinfo/processor3/stepping 3
/cpuinfo/processor3/wp yes 
/cpuinfo/processor3/siblings 4
/cpuinfo/processor3/model 4
/cpuinfo/processor3/coma_bug no
/cpuinfo/processor3/fdiv_bug no
/time 1311387215

It can also respond with JSON via .json file extension or setting the Accept: application/json header.

scott@beatbox:~% curl http://localhost:5000/cpuinfo.json
{"cpuinfo": {"processor2": {"fpu": "yes", "f00f_bug": "no", "cache_alignment": "64", "vendor_id": "AuthenticAMD", "flags": "fpu", "bogomips": "6384", "hlt_bug": "no", "apicid": "2", "fpu_exception": "yes", "stepping": "3", "wp": "yes", "siblings": "4", "model": "4", "coma_bug": "no", "fdiv_bug": "no"}, "processor3": {"fpu": "yes", "f00f_bug": "no", "cache_alignment": "64", "vendor_id": "AuthenticAMD", "flags": "fpu", "bogomips": "6384", "hlt_bug": "no", "apicid": "3", "fpu_exception": "yes", "stepping": "3", "wp": "yes", "siblings": "4", "model": "4", "coma_bug": "no", "fdiv_bug": "no"}, "processor0": {"fpu": "yes", "f00f_bug": "no", "cache_alignment": "64", "vendor_id": "AuthenticAMD", "flags": "fpu", "bogomips": "6382", "hlt_bug": "no", "apicid": "0", "fpu_exception": "yes", "stepping": "3", "wp": "yes", "siblings": "4", "model": "4", "coma_bug": "no", "fdiv_bug": "no"}, "processor1": {"fpu": "yes", "f00f_bug": "no", "cache_alignment": "64", "vendor_id": "AuthenticAMD", "flags": "fpu", "bogomips": "6384", "hlt_bug": "no", "apicid": "1", "fpu_exception": "yes", "stepping": "3", "wp": "yes", "siblings": "4", "model": "4", "coma_bug": "no", "fdiv_bug": "no"}}, "time": "1311389934"}


pip install porkchop


git clone git://
cd porkchop
git submodule init && git submodule update
python install

Extra Packages

to use the redis or thoonk plugins, you must have redis>=2.4.10 installed.

Running Porkchop


Writing Plugins

It's pretty easy to write a new plugin. They're just Python modules with some common attributes:

  • A plugin must subclass porkchop.plugin.PorkchopPlugin.
  • The plugin's class must contain a method called get_data that returns a dictionary of the information to be displayed.
  • The plugin's metric name will be derived from the module name (N.B. this means you'll only be able to define one plugin per module). To override the metric name, you may set the magic __metric_name__ property on the plugin.

By default, a plugin's get_data method will only be called if the data is more then 60 seconds old. This can be changed on a per-plugin basis by setting self.refresh in the class's ___init___ method.

These plugins can be placed in any directory you choose, and loaded by passing the -d [dir] option to porkchop.