Derived Metrics is the way of formulating a new information from existing one by performing some calculations on existing data and this process needs to be automated using functions like Average, Sum, Max, Min etc. and some unary operators along with rollup.
Users can create a new metric using metrics available in the system. For example, you can subtract the values of one metric from the values of another metric. If any metric shows the dollar revenue for a particular region, users can create a derived metric to view the same data in millions, defined as Dollar revenue/1,000,000.

Derived Metric Details

• Derived Metric Name - It is the name for newly created metrics.
  For example, if you added two metrics and one expression with the Derived Metric name - AvailMem.
  A - SysStat.FreeMem.[Tier=*;Host=*]
  B - SysStat.CacheMem.[Tier=*:Host=*]
  Expression - A + B
  Then it will create one more metrics which will be the sum of FreeMem and CacheMem
  There are some cases in which users should not specify the metric name, as the metric name will be picked from the metric name.
• Group Name - It is the name of the derived metric group. By default, it is auto-filled with the same name as Metric Group.
• Description - It is the purpose of making these derived metrics.

Metric Expression

You can add metrics, apply functions on to it, apply unary operations, apply aggregate functions like 'Group By', and can execute any expression on selected metrics.

• Metric Group - Select the Metric Group from the list.
• Metric Name - Select the corresponding metric name from the list.
• Subject - To select all/specific subjects. Further, on selecting 'Specified', the Specified and Advanced options are displayed.
  You can select the specified graphs from the Specified section.
  You can add some advanced options for adding graphs from the Advanced section.
• Function - Select which function needs to be applied on the selected metric.
  • Sum: Addition of values of all metrics specified.
  • Avg: Average of values of all metrics specified.
  • Min: Minimum value of all metrics specified.
  • Max: Maximum value of all metrics specified.
  • Sum Count: Cumulative sum of all the count values.
  • Count: Adds up count values of all the samples and creates a single metric of counts. Its behaviour better matches with the name SumCount.
• Unary Operators - You can apply one or more unary operators. A unary operation is an operation with only one operand, i.e. a single input.
  • Log2: log2 computes a base 2 logarithm.
  • Log10: log10 computes a base 10 logarithm.
  • SampleDiffPct: It is to find the % difference. previous sample value - current sample value. The diff will be absolute in %.
  • % Away from Tier Average: It is to find the % difference of the tier level average data of all hierarchy children with respect to the applied metric data It will show the metric %age away from the Tier Average.
• Adding Metrics - Add the selected metrics with parameters in the list. Each metric line has auto initials like A, B, C etc. You can add more metric rows. In the given text area, you can apply any expression on the above defined metrics like A + B or A + B /2 etc.

Formula

It contains the formula of 2 or more components . User add the Component A and after that user add the component B then inside the formula/expression input will concatenate with + operator (A+B) User can changed if he wants .

Group By Aggregation

It is a way to treat every hierarchy level independently and work like a relational "Group By" concept and hence this aggregation is named as "Group By" aggregation. This mechanism is provided to take care of some special cases where hierarchy doesn't make much sense and the user would like to see the data in a different hierarchy order.

Functions

It contains two functions - Average and Sum.

Hierarchical Metadata Component

It contains each hierarchical metadata component like Tier, Server, which is associated with that specific metric.
Note: 'Group By Aggregation', 'Functions', and 'Hierarchical Metadata Component' options are only enabled if you select 'None' in any of the metrics expressions.