This tutorial describes how and why you might need to group metrics.
The need to group metrics becomes especially relevant when you have hundreds or even thousands of metrics to analyze, which can happen when your metric expression is fairly generic or includes a widely used measure or dimension; instead of trawling through endless metrics, group the metrics according to specific dimensions or measures.
In turn, this will reduce the number of metrics displayed by grouping the relevant metrics together for each dimension or measure; a list of individual metrics that might be in the thousands is now likely to be grouped into far smaller numbers. For more details, read on below.
This tutorial consists of the following steps:
- Step 1: Define your metric expression
- Step 2: Choose the GroupBy function to group the metrics according to a specific measure/dimension
Step 1: Define your metric expression
Create your metric expression by following the basic instructions here.
OR
Click on one of the options listed under the Quick Start section when accessing the Metric Explorer. Note that if you are new to Anodot you might not have anything displayed, since the metrics in this section are previously used metrics.
For example, click BlendedCost from the displayed options, and then select a stream. You could then select individual dimensions, such as LineItem_Operation to view the metrics for this dimension. Alternatively, to save time, you can select all the dimensions (in this example, all operations), and aggregate all the metrics by using the Group By function, as described in Step 2.
Step 2: Choose the GroupBy function to group the metrics according to a specific measure/dimension
The Group By function enables you to aggregate all the metrics for each dimension, reducing by far the number of individual metrics, and providing a quick glance analysis of all the relevant metrics.
- Click Add a Function above the expression.
Note that if you want the dimension available in the description of the alert as part of the trigger, select All(*) of the dimensions. - From the displayed list of functions, select Group By.
- In the displayed list of dimensions, select the relevant dimension(s).
Note that, by default, the Sum function is selected; to change this function, click on the dropdown arrow alongside Sum and select the relevant function. See Metric Function List and Availability for more information about the available functions. - In the displayed list, define the view that best works for your analysis (see Tutorial 1: All you need to know about displaying metrics for information on the various methods of displaying metrics). For example, you can select a different time frame, or sort the metrics according to the Highest average rather than the default Alphanumeric.
You can then use the displayed metrics to analyze the data, including any spikes, and in turn, create an alert to monitor any future anomalies in the specific metrics you are analyzing. For more information, see Tutorial 3: How to investigate metric spikes.