Monitoring Log Server Process

[CFT6Server]

I am interested in putting in service checks for Nagios Log Server. Which should look at the elasticsearch and logstash services. Since those are Java processes, any suggestions on what check to see? I don't think the default Linux SNMP process check can do this due the "java" that's actually running. We are also doing agentless checks, so we are using SNMP checks.

[rkennedy]

For reference - can you provide a sample of how you're currently checking if a process is running over SNMP?

[CFT6Server]

I would be using the check_xi_service_snmp_linux_process check command. Then the $ARG1$ will be using -n 'process_name'. I've taken this from the linux snmp wizard and just modify it for the process name.

[tmcdonald]

Do you have NRPE installed on the LS server? If so, a simple bash script to check the output of service elasticsearch status and service logstash status would probably suffice.

[WillemDH]

I can recommend using this plugin to monitor a combination of a given process:
- count
- % memory usage
- % cpu usage

https://github.com/willemdh/check_lin_process

Download the latest version on GitHub and use -A to average the CPU.

Code: Select all

COMMAND: /usr/local/nagios/libexec/check_nrpe -H nlsnode -c check_lin_process -a '-p java -w 70 -c 80 -A'
OUTPUT: OK: java {CPU: 42%}{Memory: 71%}{Count: 4}} | java_cpu=42 java_mem=71 java_count=4

I'm using it for the java and the httpd process.

Grtz

[rkennedy]

Much wow, nice plugin @WillemDH!

@CFT6Server - let us know if this works for you.

[CFT6Server]

We are trying to stay away from agents as we are monitoring all devices in the enterprise without agents. This is what I was feared, but perhaps we cannot get this via just snmp and must use the NRPE client...

[hsmith]

What about check_by_ssh with a simple bash script?