Monitoring RabbitMQ with Zenoss

RabbitMQ was recently deployed in the company I currently work for. At the last minute ( as always ) they came to me and ask me to please add RabbitMQ monitoring to Zenoss. They said here is the url and now please monitor for a few stats ( using the RabbitMQ Status Plugin ). So I said to myself, I could easily just write a quick shell script to get the 3 stats that they needed and add them into Zenoss. After thinking about it…. In the near future they might ask for more than those 3 stats. So I decided to write a quick python script ( Zenoss Compatible ) to get all the stats from that status page and input them into Zenoss…

Prerequisites

1. Zenoss 2.5 and above ( I have not tested on 3.+ or <2 .4 )
2. lxml Python module    ”easy_install lxml” as the zenoss user
3. “Zenoss_Template_Manager.py” Optional, Only needed if you do not want to add all the datapoints manually… I DON’T!!!!
4. RabbitMQ Installed
5. RabbitMQ Status Plugin from http://www.lshift.net/blog/2009/11/30/introducing-rabbitmq-status-plugin
6. “check_rabbitmq.py“

Zenoss Template Manager = Download

CheckRabbitMQ.py = Download

Once you have all the above, we are ready to go..

1. copy both Python Scripts above into the /opt/zenoss/libexec/ folder  ( If you are using RedHat/CentOS ) and make sure they are executable.

As the Zenoss user run the script ….
/opt/zenoss/libexec/check_rabbitmq.py -u ‘http://rabbitmq-server:55672′ -a ‘mon-user mon-passwd’ |sed -re “s/^OK|/ /g” |sed -re “s/([A-Za-z0-9_.]+*)?=[0-9]+/-p “1,G”/g” |xargs /opt/zenoss/libexec/Zenoss_Template_Manager.py -o “/Devices/Server/Linux/RabbitMQ” -c ‘/opt/zenoss/libexec/check_rabbitmq.py -u “http://rabbitmq-server:55672″ -a “mon-user mon-passwd”‘ –template=RabbitMQ –dsource=RabbitMQStats -V $1

If you need to know how to use the Zenoss_Template_Manager.y script, check here http://www.linuxdynasty.org/howto-add-multiple-datapoints-to-zenoss-using-the-zenoss-api.html

So let me explain what the sed statements above are doing…

• “sed -re “s/^OK|/ /g”" This sed statement is removing the OK| from the beginning of the line
• “sed -re “s/([A-Za-z0-9_.]+*)?=[0-9]+/-p “1,G”/g”
  This 1st part of the sed statement is matching any letter,number,underscore, and period, any number of times until it reaches the equal “=”  ”[A-Za-z0-9_.]+*)?=”
  The  2nd part of this statement is going to match the “=” and any number of integers after it. “[0-9]+”
  Now we need to make the substitution…. So we are going to substitute, every match with a -p,  then a space and then the 1st group match in escaped quotes, then a comma and G for GAUGE.

The 2 sed statements above will do that for every match it finds. If you were to add each datapoint by hand, it would look like this….
/opt/zenoss/libexec/Zenoss_Template_Manager.py -o “/Devices/Server/Linux//RabbitMQ” -c ‘/opt/zenoss/libexec/check_rabbitmq.py -u “http://rabbitmq-server:55672″ -a “mon-user mon-passwd”‘ –template=RabbitMQ –dsource=RabbitMQStats -V -p “queue.conversion.event.tracking_msg_unack,G” -p “connections,G” -p “erlang_processes_used,G” -p “erlang_processes_avail,G” -p “file_descriptors_used,G” -p “file_descriptors_avai,G” -p “binary_memory,G” -p “memory_used,G” -p “memory_avail,G”

I hope the above scripts will save someone time and frustration……