newrelic_procstat:

A Newrelic plugin to send detailed processes statistics to newrelic at a per process level. Allowing you to gain a detailedinsight into process footprint on a server. This plugin allows you to collect detailed metrics of processes and forward them onto Newrelic, at present the plugin collects the following metrics at a per process level:

• Disk - Megabytes read from block device - Megabytes written to block device
• VM - Memory used - Major faults - Minor faults - break down of memory usage
• CPU - voluntary context swtich count - involuntary context switch count - thread count - usr time - sys time
• Network - Network states

Still To DO:

• Write init script
• Package script (add it to pypi)
• Add argument parse:
  • to accept config file
  • Set logging level

Requirements:

Plugin requires the following packages

Python package:

• python 2.7 >
• requests
• psutil
• PyYAML

Linux packages:

• systat
• python-devel
• GCC

Once we remove the reliance on python package psutil we will no longer need gcc and python-devel

Installation:

• Clone repo
• Edit config.yml file and add your process names you want to watch

 $ git clone https://github.com/AZaugg/newrelic_procstat.git
 $ nohup python newrelic_procstat/procstat.py &

Configuration:

Edit the config.yml file and add the name of the process you want to monitor

Run it:

nohup python procstat.py &

Temporary fix while I daemonize the process

Contributing:

Pull requests are welcome. A few bits and bobs to make it a completely shippable package