Skip to main content

History

The History page shows how metrics evolved over the commits of a branch, run by run. Use it to answer "when did this get worse?" or "is this input getting better?".

Opening it​

  • From the commit list, click History in the sidebar. It opens the history of the branch you're looking at, or of the reference branch.
  • From a commit page, click the name of a run on its card: you get the history of that input on the commit's branch, already filtered on it.
  • Or go to /<project>/history/<branch>.

From the History page, the sidebar has the project's navigation: the branch's commits, Code, Integrations and Milestones.

The top bar​

  • The date range and refresh button, like on the commit list. By default, the last 3 days.
  • The batch selector: which batch to follow. CI by default.
  • A filter box, with the usual syntax: path, configuration, platform, tag, tuning parameters (key:value)...
tip

The history is most useful on the CI batch, filtered on a few inputs. Filtering on a single input (by clicking a run's name) gives you its full story.

Performance over time​

A chart with one point per commit, for the selected metric (pick another at the top left; the project's configuration chooses which ones are offered).

Performance over time on main, the last two weeks: Latency p99, a line per test, and the 12 ms target dashed. One line more than triples and comes back four commits later; hovered, its point at the jump names heavy-tail-high, and the commit below copies payloads into fixed slots.
p99 per test over the last two weeks of main, with its target line: the heavy-tail test jumps at the commit that copied payloads.
  • Breakdown per test: one line per run (input and configuration) instead of a single aggregated line. Turn it off to see the median or average over all runs, but beware: commits may not have been evaluated on the same tests.
  • Relative to start (@100): rescales each line so that its first value is 100, to compare inputs with very different values.
  • Show output files: show the files a run created instead of its visualizations, in the panel below the chart.
The same chart Relative to start: every test's p99 as a percentage of its value at the first commit, so that the 200 ms bursty test and the 2 ms steady ones share a scale. Heavy-tail-high climbs to about four times its start for four commits.
Relative to start, every line begins at 100: tests with very different latencies share one scale.

Below the chart, hover any point:

  • the hovered commit becomes the New commit, and the commit before it on the line the Reference; both are shown as commit rows, with links to their commit pages.
  • With Breakdown per test, a card shows the run's visualizations (or files), compared with the previous commit. Scroll the line to find where the output changed.
PSNR of the facade photo over six weeks of main, a point per commit, with a jump at the merge of the bilateral denoiser. Below, that commit and the one before, and their output images: only the new one outlines the leaves.
Hover a point to compare its outputs with the commit before: the PSNR jumps when the bilateral denoiser is merged, and the leaves get halos.

Click a point to freeze it as the reference, then hover other points to compare them with it. Double-click to unfreeze.

tip

To find a regression, filter on the input that regressed, keep Breakdown per test on, and hover the line from right to left until the visualization changes.

Filtered on one test, hovering along its p99 line until the card changes at the commit that copied payloads; a click freezes it as the reference.

The same chart, aggregated over runs, is shown at the top of each branch's commit list.

Metrics distribution and per-test tables​

Below the chart, for the hovered (or selected) commit compared with its reference:

  • Metrics distribution: the same summary as the commit page's Summary view, with medians, changes and box plots.
  • Metrics per-test: the Metrics Table and Metrics Diff tabs, with options to sort runs by name, id or metric.

Pick which metrics to show with the metrics selector in the tables.

Sharing a history​

The chart options (selected metric, aggregation, breakdown, relative scale, output files) and the filter are saved in the URL, so you can share the exact view. See Sharing.