Question on memory utilization in predictive diagnostics.
Why the memory utilization is showing 150% or 130% or any percentage above 100% utilized ?
The Predictive graphs indicating 150% current utilization is sourced from the Oracle advisories: the advisories are showing that PGA and/or SGA need to be sized up.
So basically, Spotlight’s Predictive Diagnostics (with respect to memory) tracks the Oracle advisories over time, taking a snapshot of the memory advisories every 10 minutes, and projects forward in time, or projects forward based on expected changes in load.
For memory, the resource ceiling is essentially the point beyond which the Oracle advisories tell you that performance will significantly degrade.
You can explore the current memory situation via Spotlight’s Tuning > Memory Management drilldown. This gives a detailed insight into the Oracle advisories, and makes recommendations on memory configuration changes. Of course, that might involve acquiring more physical memory to support the required SGA/PGA configuration.
Spotlight does not alarm based on information coming back from Oracle advisories. Predictive Diagnostics is in general meant to indicate a future alarm, based on current trends. So this situation where utilization is above 100% it’s quite possible there’s a memory problem.