What Is Changing
Starting in v0.156.0, TrueFoundry AI Gateway introduces Budget Limiting V2, available at AI Gateway → Policies → Budget Limiting v2. V2 replaces the ordered-rule model in the current Budget Limiting page with independent rules and adds several capabilities:- Independent rules. Every matching rule is evaluated and the request is blocked if any rule it touches is breached. There is no rule ordering or first-match precedence.
- Tenant and team scopes. Tenant budgets are managed by tenant admins and apply across the tenant. Team budgets can also be managed by team managers and apply to a single team.
- Multiple periods per rule. A single rule can enforce daily, weekly, monthly, quarterly, and lifetime limits at once.
- Lifetime budgets. A new “resets never” period lets you set a lifetime spend cap that never resets.
- Per-entity overrides. For per-user, per-model, per-virtual-account, or per-metadata rules, you can override the base limits for specific entities.
Why This Change
Budget Limiting V1 evaluated an ordered list of rules where the first matching rule controlled allow/block. This made it hard to reason about which limit applied when a request matched several rules, and it did not support team-scoped administration, multiple periods per rule, lifetime budgets, or per-entity overrides. Budget Limiting V2 addresses these gaps:- Predictable evaluation — every matching rule must allow a request, so limits compose without depending on order.
- Delegated administration — team managers can manage their own team’s budgets without tenant-admin access.
- Richer limits — multiple periods, lifetime caps, and per-entity overrides on a single rule.
What You Need to Do
Start creating new rules in V2
Go to AI Gateway → Policies → Budget Limiting v2 and create new budget rules there. See Budget Limiting for the full configuration guide.
Audit your existing V1 rules
Review each V1 rule and note its filters, limits, and whether it acted as an override or a default. In V2, rules that previously relied on ordering will all apply together.
Recreate rules in V2
Create equivalent rules in V2. Use warn-only mode first if you want to validate spend patterns before enforcing. The migration guide maps each V1 field to its V2 equivalent.
Timeline
| Milestone | Date |
|---|---|
| Budget Limiting V2 introduced | v0.156.0 |
| Budget Limiting V1 fully removed | 3 months after v0.156.0, or once all customers have migrated — whichever comes first |
References
If you have questions about the migration or run into anything unexpected, reach out to support@truefoundry.com — we’re happy to help.