Frayme can route around a degraded provider automatically. The Fallback Rules panel binds a primary provider to a secondary one in the same category; when the primary returns degraded or disconnected, calls are transparently re-routed.
Anatomy of a rule
The dialog’s fallback dropdown is filtered to providers in the same category as the primary — Frayme refuses to swap a Sanctions/AML provider for a Device Fingerprint one.
Failover triggers
A fallback activates when:
- the primary’s last health probe is
disconnected; OR
- the primary’s
success_rate < 95% over the last 5 minutes; OR
- the primary times out N+1 times in succession (N =
retryCount on the node).
Example: KYC liveness
Resold providers (Sumsub, LSEG, BigDataCorp, Crystal) don’t ship with a default fallback. Their SLA is held by Frayme; routing around them silently would mask an issue Frayme needs to see.
What happens to in-flight calls
Mid-flight calls already issued against the primary continue to be awaited. New calls — including retries on the same node — start hitting the fallback. The workflow’s audit log records both the original target and the routed target.