SLAs & SLOs¶
Venturi separates customer AI traffic from Venturi serving-plane availability. The decision-time interceptor fails open for live AI requests; the serving-plane targets govern dashboard, query, export, and administrative surfaces.
Three terms, used precisely: an SLI is the measured indicator, an SLO is Venturi’s internal objective on that indicator, and an SLA is the contracted commitment in your agreement. The objective is always at least as strict as the commitment.
What 99.9 does and does not mean¶
The 99.9% serving-plane target applies to Venturi’s query and dashboard surface inside the customer environment. It does not describe the availability of your AI traffic, because the decision-time interceptor forwards traffic even if Venturi attribution is degraded or unavailable.
Service objectives¶
| Objective | Target | Scope |
|---|---|---|
| Decision-time interceptor latency | 50 ms P99 end-to-end budget | Decision-time observation path. |
| Stage B inference timeout | 20 ms wall-clock budget | Blocking model call where present; breach falls open to the heuristic baseline. |
| Serving availability | 99.9% target | Dashboard, query API, and exports. |
| Index freshness | P99 at or below 90 seconds target | Materialized attribution index; enterprise agreements commit at or below 5 minutes. |
| Attribution coverage | At or above 0.95 target | Share of capturable usage resolving above the confidence floor; enterprise agreements commit at 0.95, standard at 0.90. |
| Reconciliation latency | At or below 24 hours target | Reconciled-actual spend; the billable basis at coper ≥ 0.80. |
| Durability | RPO at or below 15 minutes; RTO at or below 1 hour target | Tenant attribution stores and event logs. |
Measurement basis¶
Every target is measured with a named service-level indicator at a defined point (the interceptor budget end to end at the interceptor seam; freshness at the materialized index; availability over a 28-day rolling window). Planned maintenance inside an announced window and failures of customer-side prerequisites are excluded from the availability SLI. Per-tenant objectives hold independently: one tenant’s burst is throttled before a neighbor’s experience degrades.
The error budget¶
The 99.9% objective implies a 0.1% error budget per 28-day window (about 40 minutes). Fast and slow burn-rate alerts page the on-call before the budget is gone, and an exhausted budget freezes feature rollouts and model promotions until reliability is restored: reliability spend takes precedence over change.
Forward-labeled commitments¶
The 99.95% availability step-up and the 15-minute Sev-1 mean-time-to-acknowledge target are Milestone-scoped: they attach at Venturi’s first dedicated on-call coverage milestone and are labeled forward until then (see Support tiers). Customer- or tier-specific commitments beyond this page are written into the agreement that governs the deployment.