<<< VAXAXP::NOTES$:[NOTES$LIBRARY]VMSNOTES.NOTE;1 >>> -< VAX and Alpha VMS - Digital Internal Use Only >- ================================================================================ Note 882.3 Any VMS CPU watcher (like ROBOMON software)? 3 of 4 XDELTA::HOFFMAN "Steve, OpenVMS Engineering" 23 lines 17-MAY-1996 08:32 -< SMOP >- -------------------------------------------------------------------------------- This looks like it would take an hour or two to code up, based on the $getspi C source code present over in NOTED::HACKERS -- for further information, see the "$getspi" keyword over there. ($getspi is an undocumented call that retrieves system performance information; it's the interface "underneath" the MONITOR utility.) You'll need to characterize what a "90% CPU consumer" is, as you'll get a vast amount of spurious OPCOM output if you don't also specify a sufficiently long interval for the characterization. (A C call such as "system()", where a procedure and some basic "culprit" information is passed to a spawned subprocess, can be used to log the necessary information when the threshold is exceeded.) You'll also want to check for appropriate process priorities, for the "administrative suggestion" around the use of the batch queues, and for a processor that is underconfigured and/or overloaded. You'll also want to check the toolshed. See METOO::SW_TOOLS_CATALOG and METOO::TOOLSHED. And -- if you don't want to spend the time to code tool this up, and if you can't find a tool in the toolshed -- you will want to order ROBOMON or similar package.