Analytics
Every run, filterable — who ran what, on which machine, how long, and what it cost.
The Analytics page is the run-history table for your whole workspace: an infinite-scrolling, filterable log of every run.
What's in the table
- Run ID — click through to full run details (inputs, outputs, logs)
- Time — when it ran
- User — who triggered it (essential for org accounts)
- Machine & GPU — where it ran
- Status — success / failed / cancelled / timeout
- Duration — queue, cold-start, and execution time
What it's for
- "Who's spending our credits?" — filter by user, see their runs and durations; pair with the Usage page for the dollar view
- "Why is this workflow slow?" — compare durations across runs; long cold starts point at machine configuration, long execution at the workflow/GPU choice
- "What failed overnight?" — filter by status, click into logs
For per-GPU spend over time, the Usage page has the day-by-day chart; Analytics is the per-run microscope.