I don't necessarily believe any of the metrics are that sensible at an *instance* level, they are very coarsely summarized ( 5 min periods would be better for CPU but I don't believe they are there) and so both avg and max/min are meaningless (at the very least without std deviation) On 2 Sep 2010 16:40, "Rich Jesse" <rjoralist2@xxxxxxxxxxxxxxxxxxxxx> wrote: > That'll give the host cpu and not the db CPU > > I suspect that the OP is after something along th... You beat me to it, Niall! Although I'm not sure which CPU metric would be the proverbial "best", so I came up with this one to let the OP decide: SELECT mmd.* FROM sysman.mgmt$metric_daily mmd JOIN sysman.mgmt$target mt ON mmd.target_name = mt.target_name AND mmd.target_type = mt.target_type AND mmd.target_guid = mt.target_guid WHERE mmd.metric_column like '%cpu%' AND mt.target_name = :DB_NAME AND mt.target_type = 'oracle_database'; The extra view and columns I used for performance, although that's probably debatable... And, of course, this is run directly against the Grid Control repository DB and *not* the target database. (thought it worth mentioning) GL! Rich