Roll up and Aggregation Edit on GitHub Request doc changes

Roll up

Data displayed in a widget is rolled up from the underlying data points collected during acquisition. For example, if you have a line chart widget showing Storage IOPS over time, you might want to see a separate line for each of your data centers, for a quick comparison. You can choose to roll up this data in one of several ways:

  • Avg: displays each line as the average of the underlying data.

  • Max: displays each line as the maximum of the underlying data.

  • Min: displays each line as the minimum of the underlying data.

  • Sum: displays each line as the sum of the underlying data.

Steps

To set the roll up method, do the following.

  1. In your widget’s query, choose an asset type (for example, Storage) and metric (such as IOPS - Total).

  2. For Roll up, choose a roll up method (such as Avg) and select an attribute or annotation by which to roll up the data (for example, Data Center).

    The widget updates automatically and shows a line for each of your data centers.

You can also choose to roll up all of the underlying data into the chart or table. In this case, you will get a single line for each query in the widget, which will show the average, min, max or sum of the chosen metric for all of the underlying assets.

If you have set a filter for the query, the data that is rolled up is based on the filtered data.

Note that when you choose to roll up a widget by any field (for example, Model), you will still need to Filter by that field in order to properly display the data for that field on the chart or table.

Aggregating data

You can further align your time-series charts (line, area, etc.) by aggregating data points into minute, hour, or day buckets before that data is subsequently rolled up by attribute (if chosen). You can choose to aggregate data points according to their Avg, Max, Min, or Sum, or by the Last data point collected during the chosen interval. To choose an aggregation method, click on More options in the widget’s query section.

The minimum allowed interval is ten minutes. A small interval combined with a long time range may result in an "Aggregation interval resulted in too many data points." warning. You might see this if you have a small interval and increase the dashboard time frame to 7 days. In this case, Insight will temporarily increase the aggregation interval to 1 hour until you select a smaller time frame.

You can also aggregate data in the bar chart widget and single-value widget.

Most asset counters aggregate to Avg by default. Some counters aggregate to Max, Min, or Sum by default. For example, port errors aggregate to Sum by default, where storage IOPS aggregate to Avg.