Specifications

Chapter 6: Business views
Capacity planning
All lines on the performance charts include an optional "best fit" line that shows a
flat line which best fits the points for that metric. To see the best fit line, select the
Best Fit option at the top of the chart.
The best fit line is based on a classic least squares linear regression calculation
of the data points. It is only meaningful if there is an underlying long term trend in
resource utilization. For some resources, for example, disk space, this is almost
inevitable. For others, for example, CPU, it only applies if the externally applied
system load is driving a trend in the data set. The calculation is done on the data
range selected in the graph. Hover over the best fit line to see the following
values:
l Gradient is the slope of the line, it displays the rate of growth of the
performance metric or trend.
l r2 is the r-squared coefficient of determination. The range of possible values
is 0.0 to 1.0. A higher value means that the data better fits a straight line
projection with growth over time. Do not lend too much weight to this value if
there is only a small number of data points in the graph. For example, if
there are only two data points in the graph, the value of r2 will always be
1.0. A low value of r2 indicates either a lot of variability in the data, or that
the data do not change significantly with time. Either of these cases means
that extrapolating a "drop dead" date on the specific resource is invalid.
For any metric that is a percentage of total resources, the best fit line also
includes an estimate for when that value will reach 100%. This gives you a
predictive analysis of how your resources are trending over time, so that you can
plan when to insert more resources.
Overlays
At the top of the Performance Analyzer chart there are other options that control
data that can be overlaid on top of the time line in context of the lines being
displayed.
134