High CPU Utilization issues when importing the DHCP Management Pack

Some of us experienced a very high CPU Utilization issue after importing the DHCP Management Pack (6.0.6383.0). It was due to some heavy usage of discovery scripts.

I'm glad that its now all fixed with the release of the new 6.0.6452.0 version that is now in the MP Catalog

Here are some of the changes that was made by the product team. (Thanks guys!! sending you the some beer vouchers soon)

• Changed the discovery script so that it discovers same instances every time and does not cause changes in their properties (it causes changes in OpsMgr configuration each time).
• Rewrote script-based module in the following way – now it returns all data for all components (scopes, superscopes, etc.) instead of returning data only for requested one. It allowed SCOM to cook it down and now instead of 1000 scripts only one instance is run.
• Rewrote script-based module in the following way – instead of returning many property bags (4000 for 1000 objects) it returns only one big property bag. It allowed to lower significantly transferred data.
• Added performance mapper for unit monitors based on huge property bags. It allowed to significantly improve performance since now only small package of data is transferred between MonitoringHost and HealthService.
• Removed On-Demand-Detection for monitors based on huge property bags since SCOM SP1 does not support cook down for such scenarios and it leads to very high CPU consumption upon any changes in configurations/restarting HealthService.
• Optimized script populating all data from DHCP server (it analyzed with regex expression big output upon iterating through all objects every time, now it does it only once). It allowed to significantly improve performance for this module (it consumed 100% of one CPU for minutes).