Quota Cost Control is the per-tenant cap and budget page. AI spend ceilings (which feed Cost Circuit Breaker), query-cost budgets (which feed Query Hub), storage quotas, retention policies, alert thresholds for spend. One page that ensures tenants do not spend you out of business and CFO has the numbers they expect.
Four families: AI spend (LLM tokens), query spend (BigQuery / Snowflake), storage (incident archive, ledger, bundles), retention windows (how far back data lives). Each family has soft and hard caps per tenant. Defaults come from the plan tier; overrides are first-class config the tenant or the operator can adjust.
Soft cap fires a notification (email, Slack) so the tenant can plan. Hard cap blocks the over-quota action (Cost Circuit Breaker trip, query reject, storage write reject). Soft is set by default at 80% of hard; both are configurable per tenant. Crossing thresholds is logged for the monthly bill review.
Each plan tier ships with default caps appropriate for the price. Basic has small caps suitable for kicking the tires. Enterprise has caps tuned to actual production usage. Tenants on a usage-based contract see their own custom caps. Caps move with the plan tier automatically when a tenant upgrades or downgrades.
A monthly bill with attribution: which agents, which queries, which storage, which users drove the spend. The bill exports as PDF and CSV. Surprising line items have explanations attached automatically: "30% of AI spend was inc-4821 in April." Finance reviews are easier when the bill explains itself.
Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.
A runaway tenant is a real cost; a hard-capped tenant is a happier customer. Quota Cost Control draws both lines.