Exploring heatmaps

This page describes the features of the Key Visualizer interface.

Before you read this page, you should be familiar with the overview of Key Visualizer.

You may also want to:

Adjusting brightness

Key Visualizer heatmaps use dark colors to indicate normal or low values, and bright colors to indicate high values. If the colors in a heatmap are too bright or too dark, it might be difficult to see important patterns for a metric.

Values associated with each color are defined in the legend to the left of the Adjust Brightness buttons.

Keep in mind that for most metrics, Key Visualizer identifies which values are "high" and "low" based entirely on your usage of that database. For example, if you perform 20 reads per second in a specific row range and 1 read per second in other row ranges, Key Visualizer will use a bright color for that specific row range, even though 20 reads per second is a low number for Cloud Spanner.

To adjust the colors that are used for different values, use the + / - buttons on either side of the Adjust Brightness icon. When you increase the brightness, you lower the range of values that each color represents; when you lower the brightness, you increase the range of values.

Zooming in on time periods and row key ranges

When you open a heatmap, it displays values for the entire time period included in the scan, as well as the entire range of row keys in your table. If you notice an issue during a specific time period, or in a specific range of row keys, you can zoom in to examine the issue more closely.

There are several ways to zoom in or out:

  • Scroll up or down within the heatmap.
  • Click the Rectangular Zoom button, then click and drag to select an area to enlarge.
  • Press and hold the Shift key, then click and drag to select an area to enlarge.
  • Return to the previous area of the heatmap by pressing the Backspace or Delete key.
  • Zoom out to view the entire table by clicking the Reset Zoom button or pressing the R key.

Pinning details about a metric

As you move your cursor over a heatmap, Key Visualizer displays a tooltip that contains details about the metric that is under the cursor. You can pin this tooltip in place by clicking on the heatmap. Pinned tooltips show more detail about the metric. Pinning a tooltip also makes it easier to take a screenshot of the tooltip or copy information from the tooltip.

Switching metrics

You can choose the metric to display in the heatmap from the drop-down metric list located at the top left of the page.

For descriptions of each metric, see Key Visualizer Metrics.

Finding connections between different metrics

You can view multiple metrics at the same time in Key Visualizer when you want to compare metrics or find correlations between metrics.

To view multiple metrics and their values at the same time, click Show info panel near the top right corner. A list of metrics appears in a pane to the right of the heatmap.

Example showing the multimetric pane for_a_scan

To examine several metrics at once, use the following tools in the info panel:

  • To show or hide miniature heatmaps for all metrics, click Expand All or Collapse All at the top right.
  • To show an individual metric's heatmap, click the metric name.
  • To hide an individual metric's heatmap, click the metric name again.
  • To display a metric's heatmap in the main window, click the solid arrow at the left edge of the metric's title bar.

    Left arrow shows a metric in the main window

  • Hover over a miniature heatmap to see corresponding activity in the main view.

To compare different metrics for a row range at the same point in time, you can pin a value in a scan, then switch to a different metric at the same point in time:

  1. Point to a position of interest on the large heatmap in the main window.
  2. Click to pin the tooltip. The values in the side panel stay fixed to the row range and point in time you've pinned.
  3. To display a metric in the large heatmap, click the solid arrow on the left for that metric. The tooltip in the main window shows you the full details for that metric.

What's next