Docs center

Self-hosted docs for Guardrails for AI.

Technical guidance for observability, guardrails, permissioning, and automation in one first-party documentation surface.

Self-hosted docs

Failover Patterns

Keep your SLA commitments when providers degrade by routing traffic to approved alternatives.

docs.cyiro.comproduction-ready guidance

Define allowlisted endpoints for each route so reroutes remain compliant with policy and security constraints.

Example: Allowlist fallback endpoints in failover policy configuration
Example: Fallback endpoints must pass same policy checks as primary
Example: Security constraints applied consistently across all destinations
Request: POST /v1/failover/policies {route_id: "prod-chat", fallback_endpoints: ["backup-chat-v1", "backup-chat-v2"]}
Response: {policy_id: "fail-123", status: "active", approved_endpoints: ["backup-chat-v1", "backup-chat-v2"]}
API Reference: See [Failover Policies](/docs/api#failover-policies) for full contract details

Use correlated runtime and status signals to trigger failover at the right moment, not too early or too late.

Example: Correlated latency spikes and provider status changes trigger failover
Example: Multiple degradation signals required for high-confidence triggers
Example: Gradual rollback when primary provider recovers
Request: POST /v1/failover/triggers {route_id: "prod-chat", signals: ["latency_spike", "provider_outage"], confidence: 0.95}
Response: {trigger_id: "trig-456", action: "activate_failover", timestamp: "2024-03-15T14:40:00Z"}
API Reference: See [Failover Triggers](/docs/api#failover-triggers) for full contract details