Ad Widget

Collapse

Finding a single runaway process from a list of many

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • dthacker
    Member
    • Feb 2014
    • 42

    #1

    Finding a single runaway process from a list of many

    Problem: I have many proceses like this
    Code:
    oracle   28735  0.0  0.2  69928 26848 ?        S    10:08   0:01 /u01/app/oracle/product/mid1012prd/bin/frmweb server webfile=HTTP-0,0,0,rms13prd,10.53.12.174
    . Sometimes one of the processes goes into a runaway status and borks the server. I have been asked to detect any of process that has accumulated more that 500 minutes of CPU and trigger.

    My first thought was to use UserParameter. I set an awk command that pulls just the accumulated CPU and added to UserAgent using the tutorial in the Zabbix docs.

    Code:
    UserParameter=oracle.webfrm,ps -axu | grep '[f]rmweb' | awk '{print $10}
    When I run the command in zabbix agent I hit this error:
    Code:
    [root@prdrmsappl01 zabbix]# zabbix_agent -t oracle.webfrm
    oracle.webfrm                                 [m|ZBX_NOTSUPPORTED]
    How can I resolve that?

    TIA Dave
Working...