Best practices

Operational practices that recur across production Agent Router Enterprise deployments: resilience, cost control, safety, and configuration management. Each card links to the guide that covers it in full.

---

Resilience and routing

Improve resilience with fallbacks

Define ordered fallback policies so requests survive a provider outage.

Plan high availability and disaster recovery

Size the deployment and plan recovery for a production gateway.

Cost and governance

Working with budgets

Track spend against limits with opt-in enforcing actions; hourly token ceilings are the separate rate-limit control.

Reduce cost with traffic splitting

Distribute traffic across backends by weight to manage cost.

Safety and configuration

Protect requests with guardrails

Apply inline content filtering to inbound and outbound traffic.

Guardrails best practices & FAQ

Roll out, tune, and operate guardrails without outages; investigate blocks and understand scope.

Manage configuration as code

Version Agent Router configuration and apply it through automation.