Monitor Process Count/Availability (Sample Plug-In Monitor)

[rajib]

This is a fully functional plug-in monitor that can be used to monitor the availability of a process (whether it is running) on a remote server. It can also monitor the number of instances of a particular process. The monitor illustrates the two different types of information that can be collected by Traverse - state (up/down, running/not-running, etc) information represented by a set of discrete values and quantitative where the polled value is compared against a set of configured thresholds.

In order to run this script, extract the contents of the attached archive directly under Traverse installation directory and restart the Web Application and Data Gathering Engine components. If there are multiple Traverse servers, this should be performed on all BVE/DGE servers. The plug-in will appear as a new "hostres" monitor under Administration -> Devices -> Tests -> Create New Standard Tests. On the next screen, select one or both types of tests to run. Finally, provide the name of the process that is normally visible in the output of "ps" command (Linux/Solaris) or Task Manager (Windows). The monitor uses SNMP to collect the necessary information and therefore requires that at least one SNMP test is configured.

Standard corporate disclaimer applicable.

[cbell]

Can this be used to poll data that would also be available using the 'top' command in Linux? More specifically on ESX hosts using the 'esxtop' command?

[rajib]

Hi Chris,

The posted script is designed for the specific purpose of counting number of instances of a process, but the concept is applicable for monitoring any metric. The plug-in framework can be used to create a new monitor that (for example) logs into an ESX server via SSH, executes "esxtop" command with certain parameters and retrieves memory/CPU utilization for a given VM. You have complete control over what operations you would like to perform or how the collected data is processed within the plug-in monitor. You simply need to provide Traverse with a processed numerical result that will undergo normal processing - compare against thresholds to determine test status, trigger action profiles, aggregate historical results, etc. The developer's guide (http://zyrion.com/support/docs/)contains detailed information regarding the plug-in framework and other advanced capabilities.

[it-drift]

Hi Rajib,

I use the hostres plugin to monitor processes on servers, but I have a server where the executables are located on the D-drive in some folder, and Traverse is showing me that the processes are not running. Can the location of the executables, be the reason for this?

Process names from pslist.exe:, and all are running.
CollectionGuardian 1820 8 9 269 21780 0:00:15.328 576:34:01.758
CollectorService 1864 8 16 529 30868 0:06:03.843 576:34:01.695
GenericTaskService 1932 8 14 425 28928 0:02:58.593 576:34:01.602
ProcessorService 2004 8 19 473 109172 0:57:30.484 576:34:01.555
ReportingService 2064 8 11 357 26712 0:03:07.671 576:34:01.477
PhonEXSQLService 2112 8 10 383 25876 0:01:15.062 576:34:01.336

Any workaround for this problem?