feature request...tweak timeouts

[dougsyer]

Today on the old zenpack I have a fairly complex set of transforms that does counting escalation based on type of WMI error and count.

I suspect the timeouts will be less now with this zp but one thing I realized is that without the timeout/limiter(ie ease off on collection like max oids) zproperties you have with zencommand or zenperfsnmp you are forced to expose your system to huge bursts of load due to waves of counting transforms going off when there are connectivity issues or windows issues that prevent the processing of monitoring data..

If you have timeouts / limiters you can push that work away from the core and onto the collectors (where it should be) and you can customize them per device/site etc based on connectivity /health of the system if you need to.

If I remove the /status/wmi count transforms (and we have /command ones but they dont get hit nearly as much because wmi is more of a challenge) ...the issue is that our NOC gets exposed to alot of non/actionable events. which eventually they become numb to.

for more detail see parature case # 27296