Docs »

Workspaces »

Dashboards »

Widgets »

Chart Metrics Explorer

The Chart: Metrics Explorer widget is an interactive chart builder for metrics.

Unlike the other chart widgets, you don't write a data query first. Build the chart by adding series and adjusting them in place, watching the result update as you go – then export it once it looks right.

It's a card widget, available on metric records and on other record types.

Building a chart

The widget opens with five series already defined against the current record – Samples, Sum, Average, Min, and Max – but only Sum is visible, so a first open looks like a single line. The rest are waiting behind their visibility toggles.

Each series is configured independently:

Control Notes
Metric Which metric to plot
Function How to aggregate its samples
Dimension filters Narrow the series by dimension, matching (is) or excluding (not)
Type Line, bar, or area
Axis Plot against the left or right axis
Stacking Stack this series with others
Color The series color
Visibility Show or hide without deleting it

The chart's range and period apply to every series.

Functions

count, sum, avg, min, max, distinct, faceted_average, faceted_min, faceted_max

The faceted_* variants aggregate across a metric's dimension rows within each bin – a sum of minimums, for instance, rather than the minimum of the whole bin.

Periods

minute, hour, day

Pick the period to match the range: a day-level period over a 24-hour range gives you one bar.

Exporting

Nothing you build in the explorer is saved. Series, range, period, chart type, axis, color, and visibility all live in the browser, and closing the card discards them. Exporting is the only way to keep a chart you want back.

Once a chart is right, it can leave the explorer three ways:

Export Use it for
metrics.timeseries data query Reusing the query anywhere data queries are accepted
Chart KATA widget Putting the finished chart on a dashboard
Explorer config Saving the widget's own configuration

This makes the explorer a practical way to write a metrics.timeseries query: build it visually, then export the query rather than composing it by hand.

Related