AI usage
Monitor your organization's monthly token budget, hourly rate limits, and skipped evaluations.
Reviewed AdminVertiqa 1.54+
The AI usage page shows what your goals are spending in agent compute and where Vertiqa had to throttle. It's read-only — a dashboard, not a control panel.
What the page shows
- Monthly token budget — total tokens available this month and the percentage you've consumed. Token budget is per-organization.
- Hourly evaluation rate limit — how many goal evaluations the org is making against its per-hour ceiling.
- Skipped evaluations — evaluations Vertiqa chose not to run, broken
down by reason:
- Budget exceeded (red) — you hit the monthly token budget.
- Rate limited (yellow) — you hit the hourly cap.
- No matches (emerald) — the goal evaluated but nothing qualified.
- Other system reasons.
How to use this page
- Watch the budget bar through the month. If it's racing ahead of the calendar, the most expensive goals are candidates to tighten.
- A spike in rate limited skips means a sweep is competing with ongoing evaluations. Stagger sweeps or scope them tighter.
- A high no matches ratio is fine — that's the goal's filter doing its job.
Tip
If you're consistently hitting the monthly budget, the fix is usually a narrower filter on one goal, not a budget increase. Tightening the set of records a goal evaluates is the cheapest lever.