Main pages
- Compute: every registered compute, its state, type, and last-seen time. Click a row for the per-compute drill-down.
- Executions: every workload that ran on every compute, with status, resource usage (cores, memory actual vs requested, GPU), and runtime. Click a row for the execution detail page.
- Waste (owners and admins only): aggregated waste across all executions: per user, per team, per compute.
- Teams: invite colleagues, control which computes each member sees, manage roles. See Teams & roles for what each role unlocks.
- Settings → API Keys: manage your personal
exp_user_key.
Connecting a new compute
Use the CLI:Intelligence
Completed runs show a generated run summary on the execution detail page. For failed runs, the console gives you a one-clickexpanse diagnose <execution-id> command; run it from the CLI to get the root cause, suggested fix, and cited evidence. Resource recommendations come from expanse analyse in the CLI.