Analytics

Every Run you upload feeds the analytics surface, which is built into the dashboard, honors the same role and team scoping as the rest of TofuPilot, and works the same way on managed and self-hosted deployments.

What analytics covers

Analytics groups every metric into four families, so you can find what you need from a single tab.

Reference Samples, which are Units flagged as golden or failing, are excluded by default. You can toggle them back in from the filter bar.

Runs

The runs chart shows daily run volume grouped by outcome: PASS, FAIL, ERROR, TIMEOUT, and ABORTED.

Units

The units chart shows the daily count of unique units you tested. When runs exceed units, units have been retested, and the gap is a useful proxy for rework load.

Run duration

The run duration chart shows the daily average wall-clock duration of a run. For a per-phase breakdown, see Phase durations.

Filters

Every chart and table on the analytics tab share the same filter bar, so a single change updates the whole view.

• Date range: pick a preset or define a custom window.
• Source: choose user runs from CLI or production runs from a Station.
• Procedure version: compare yield before and after a Deployment.
• Part name and Revision: slice yield by revision.
• Batch: slice yield by production cohort.
• Reference Samples: toggle golden and failing units back in.

Filters apply to every chart on the tab at once, so you never see a mixed view.

How metrics are computed

A few rules govern how every chart in the analytics surface counts your data.

• Reference samples are excluded from yield, volume, and duration.
• ABORTED runs do not count toward yield.
• Yield is computed per unit per day, by attempt. A unit you test three times produces one FPY, one SPY, and one TPY data point. See Yield for the full breakdown.