This guide explains how to effectively search and filter multiple performance graphs within the Browser Performance Infrastructure. While the Explore (/e) view is designed for deep-diving into a single, comprehensive graph, Multigraph (/m) is optimized for viewing and synchronizing multiple metrics simultaneously.
Note: All examples and links in this guide use the Chrome Perf instance (https://chrome-perf.corp.goog/m/), but the concepts and instructions apply to all Performance Dashboard instances.
You can access the Multigraph view in two ways:
/m path (e.g., https://chrome-perf.corp.goog/m/).Note: This is the default view for most dashboard instances.
The Test Picker is the primary tool for building queries to find specific metrics. It operates on a dynamic key-value hierarchy:
benchmark, bot, test). When you select a key, you are prompted to select one or more matching values.benchmark, bot, test) is determined by the include_params setting in the instance configuration. This allows the search hierarchy to be customized based on how the data is indexed (see example configuration). Keys not specified in the hierarchy are not set by default, but can be assigned default values via default_param_selections (see example).Multigraph operates in Sync Mode by default. In this mode, the graphs are tightly coupled to the Test Picker. This is optimized for rapid exploration and comparative analysis: as you adjust filters in the Test Picker, the graphs update instantly, allowing you to quickly correlate performance changes across different metrics or bot configurations without manual effort.
To plot the initial graph, select the desired parameters and click the Plot button in the Test Picker. This will plot a single graph with the selected query.
You can add a trace by selecting a value in the Test Picker that narrows your query down to a unique trace. It will automatically be plotted on the chart.
Selecting an additional value in the same field...
...will instantly plot the new trace on the same chart alongside the first one.
You can remove a trace by de-selecting its value in the Test Picker. This immediately removes the corresponding trace from the chart.
When you want to quickly add all available values for a specific field to your chart, you can check the “All” box next to that field in the Test Picker.
This will simultaneously plot all matching traces on the chart.
You can create multiple synchronized graphs based on a specific parameter by toggling the “Split” button next to any key in the Test Picker.
For example, splitting by bot will generate a separate graph for each bot matching your query.
The separate graphs will remain synchronized as you zoom or pan across the timeline.
Independent Graph Mode lets you build custom dashboards with completely unrelated metrics. It is currently enabled by adding ?manual_plot_mode=true to the URL, but can also be set as the default via a configuration change. Note that in this mode, selections in the Test Picker will not affect the displayed graph unless the Plot button is clicked, which will render a new graph.
To plot a new graph, select the desired parameters and click the Plot button in the Test Picker.
This will generate a new chart at the top of the page with the traces matching your current query.
Each Plot click generates a new graph. For example, selecting a different query and clicking Plot again...
...will render a completely separate, independent graph above the first one.
If you have a trace that you want to use as a starting point for a new search, you can click the Query Highlighted button on any chart. This will backfill the Test Picker with all the keys and values that define that specific trace.
To remove an independent graph, click the X button located in the top-right corner of the chart container.