What are appropriate alarm thresholds for veritas cluster server

Hi:

- We have developed our own agent for monitoring vcs,
  but we are uncertain about appropriate alarm thresholds

  - That is, when should a warning alarm be raised
    instead of an error alarm ?

  - Could someone please look at our alarm thresholds, and
    provide some feedback ?

Thanks
John



  -1)  when a VCS Cluster's state is not equal to "RUNNING",
         an error alarm is raised

  - 2) when the state of a System in VCS is not equal to "RUNNING",
        an error alarm is raised

  - 3) when a Group's Probed valued is not equal to "Y",
         a warning alarm is raised

  - 4) when a Group's AutoDisabled value is not equal to "N",
        a warning alarm is raised

  - 5) when a Group's State matches "FAULTED" (a regular expression match),
         an error alarm is raised

  - 6) when a Resource's State matches "FAULTED" (a regular expression match),
         an error alarm is raised

  - 7) when a Resource's Probed value is less than "1",
        an error alarm is raised

  - 8) when a VCS Agent's Faults value is greater than "0",
        an error alarm is raised

  - 9) when the VCS shared disk directory does not exist,
        an error alarm is raised

 -10) when VCS GAB Status is eqal to "DOWN",
        an error alarm is raised

 - 11) when one network interface for GAB is no longer "UP",
        an error alarm is raised