Data explorer

The Data explorer greatly enhances your abilities to query and chart metrics.

• Analyze multidimensional metrics, whether built in to Dynatrace or ingested from StatsD, Prometheus, Telegraf, and others channels.
• Choose a visualization type:
  • Graph
  • Stacked column
  • Stacked area
  • Pie
  • Single value
  • Table
  • Top list
  • Heatmap
  • Honeycomb
• Add your visualization right to your dashboards for easy access and sharing.
• Provides a foundation for metric calculation in charts on dashboards.

Quick start

If you prefer to skip the technical details and learn by doing, try the Data explorer quick start. You'll learn how to:

• Create a simple chart in the Data explorer
• Use the chart in the Data explorer
• Pin the chart to a dashboard as a tile

But if you want the technical details now, read on.

Query components and concepts

Every metric query is composed of multiple optional components:

Limitations

• 10 metrics maximum per chart
• 100 series per metric

Examples

Table with two metrics (split)

In this example, we select metrics CPU usage % and Memory used %, break it down by host for both, and display it as a table, so that the rows are hosts and the columns show the metric values per host.

• A: CPU usage % (builtin:host.cpu.usage), Average, Split by Host
• B: Memory used % (builtin:host.mem.usage), Average, Split by Host
• Visualization: Table

The complete query should look like this:

Example tile:

Graph with two metrics

In this example, we select the same metrics and display them as a graph.

When you set Visualization to Graph, the Visual settings are displayed, where you can select how to graph each metric. In this case, CPU usage % is an area chart (the area between 0 and the value of the metric is filled in) and Memory used % is a line chart (a single line representing the value of the metric over time).

• A: CPU usage % (builtin:host.cpu.usage), Average, Split by Host
• B: Memory used % (builtin:host.mem.usage), Average, Split by Host
• Visualization: Graph
• Visual settings:
  • A = Area
  • B = Line

The complete query should look like this:

Example tile:

Visual settings

The selection of visual settings available depends on the query and visualization type. Dynatrace offers the following visualization types:

• Graph
• Stacked column
• Stacked area
• Pie
• Single value
• Table
• Top list
• Heatmap
• Honeycomb

Entity linking

For better drilldown capabilities, we have added links to entities.

In this example, the visualization was showing a host behaving erratically. We selected it to see details about the host in a pop-up, and we can select the View host button to drill down directly to the Dynatrace page for that host.

Fold transformation

(Applies only to value visualizations such as single value, pie, top list, table, and honeycomb.)

The fold transformation combines a data points list (a timeseries: a collection of data points over the time period) into a single data point.

• The default fold transformation is Auto, which automatically selects the most appropriate time aggregation based on the metric.
• You can manually override Auto with any of the following: Last value, Average, Count, Maximum, Minimum, Sum, Median, Value, Percentile 10th, Percentile 75th, Percentile 90th.
• If you need to see the last reported value for a metric (rather than an aggregation), select Last value.

Trendline

To be able to determine if the current state is the common behavior of the time period, turn on Show trendline.

Advanced query editor

To fully utilize the power of the Metrics API v2 queries from within the web UI, switch to the Code tab.

The Code tab enables you to create metric expressions, timeframe shifts, and much more directly in the web UI and, of course, use this power to create vizualizations for your dashboards. Start by checking out metric expression examples.

For Code tab details, see Data explorer Code tab (advanced query editor).

Share your results

If you are doing analysis in the Data explorer and find some interesting results that you want to share with other authenticated users, or if you want to be able to revisit the same results yourself using a later timeframe

1. In the Data explorer, select More (…) > Share link in the upper-right corner of the Data explorer
2. Determine the timeframe to associate with the link:
   • To share the link with the current timeframe, turn on Use the current timeframe.
   • Otherwise, the shared query link will specify the current query and settings except the timeframe.
3. Select Copy to copy the link to your clipboard.
4. Share the link with any other authenticated Dynatrace user or keep a copy for your own use.

Export to CSV file

To export to a comma-separated values (CSV) file

1. In the Data explorer, select More (…) > Export to CSV in the upper-right corner.

   Note:

   • CSV export is available for all visualizations except honeycomb and single value
   • Values exported to a CSV file reflect the formatting specified with the Unit and Format settings in the Visual settings section.

2. A CSV file of the results is saved to your local machine.

The file name indicates the metrics, date, and timeframe.

For example:

• CPU usage % (May 24, 2022, 11_41 - 13_41).csv—contains results from metric CPU usage %, run on May 24, 2022, for a two-hour timeframe of 11:41-13:41.
• CPU usage % +1 (May 24, 2022, 13_19 - 13_49).csv—contains results from metric CPU usage % and one more metric, run on May 24, 2022, for a half-hour timeframe of 13:19-13:49.

Sort

By default, results are sorted in descending order based on the aggregation chosen.

To explicitly set the sort order through the query editor

1. Select the + button.
2. Select Sort by.
3. Set Sort by to Ascending or Descending.

Limit

By default, the number of metrics you see if they are split by a dimension is 10.

To explicitly set a limit through the query editor

1. Select the + button.
2. Select Limit.
3. Set Limit to 1, 10, 20, or 100.

Auto-extended filtering

Auto-extended filters leverage the Dynatrace topology (entity model) to offer additional filter dimensions not available in the original metric. They work on both the tile level and the dashboard level.

• On a tile level, select them when you set Filter by in the Data explorer by simply selecting the original dimension of a metric that has the relationship assigned. For example, a metric that captures the performance for Synthetic events has a relationship to a Synthetic monitor. Using the topology, you can first select the related Synthetic event in the filter and then, besides the name, tag, id, or health state, you also get an additional option to pick the related monitor.
• On a dashboard level, while you can't pick desired relationships, Dynatrace automatically extends the metrics where possible, so that, when you pass a dynamic filter, it can apply to a tile with that metric.

Example 1: Synthetic step metric extended by Synthetic monitor

Some performance metrics for Synthetic events lack the ability to filter them by monitor. However, the same event could happen in multiple monitors, and to look at a single monitor's performance you need the ability to filter for them.

With automatically extended filters, you can now filter on the Synthetic test step.

1. Select the metric (for example, Action duration - load action (by event) [browser monitor]). It has entity type SYNTHETIC_TEST_STEP.

2. Add the filter.

3. Select auto-extended dimension.

4. Resulting filter.

Moreover, you can now use auto-extended filters on your dashboard, so there's no need to configure multiple tiles to see the same metric for different monitors or different hosts.

5. On the Dynamic filters tab of your dashboard's Dashboard settings page, add a filter for Custom dimension.

6. Select Synthetic monitor.

7. Save your changes and display the dashboard. Now you can filter all tiles on the dashboard by Synthetic monitor. The dashboard will automatically, for each tile, check whether any such relationship exists. So every tile (without filters set on a tile level) that has a Synthetic event metric will be filtered the moment a relationship exists between the step in the tile and the monitor you picked.

Example 2: Host metric extended

Here, we extend the host metric by EC2 instance.

1. Create a host-related tile with a host metric (for example, CPU usage % - builtin:host.cpu.usage).

2. Apply a related filter such as EC2 instance (runsOn).

   Now the tile with all hosts is filtered to only the hosts running on that EC2 instance. This is possible even though the dimension EC2 instance does not exist on the original host metric. By leveraging the topology (entity model), Dynatrace can filter the hosts based on that relationship.

In this variation, the host metric is extended by host group.

1. Set a filter for Host.Host Group (isInstanceOf) and pin the tile to our dashboard.

   

2. We are now able to filter the dashboard tiles by host group.