How does failover work?
The orchestration layer watches each provider’s responses. A payment that times out or returns a technical error can be sent to the next eligible provider. When errors from one provider rise above a threshold, new traffic can be moved away from it entirely until it recovers, a pattern often called a circuit breaker. The merchant receives one final status, whichever provider completed the payment.
Why does failover matter for revenue?
Provider outages are rare but expensive, and they often happen at peak traffic. As an illustrative example: a merchant taking 3,000 payments an hour at an average of USD 30 attempts USD 90,000 an hour. A two-hour outage at its only provider leaves USD 180,000 of payments unprocessed, and some of those customers will not come back. With failover, most of that traffic moves to a second provider within seconds.
What are the risks of failover?
The main risk is duplication. If the first provider approved the payment but the response was lost, a retry on a second provider charges the customer twice. Failover also needs a backup provider that is actually eligible for the traffic, with the right countries, methods and limits, and enough capacity to absorb a sudden increase. A backup that is never used in normal operation may fail when it is needed, so some live traffic should keep flowing to it.
Where does Omno use it?
Omno’s routing and cascade rules send a failed payment to the next eligible PSP, PSP eligibility rules make sure the backup can take the traffic, webhooks report the final status once, and the routing-decision log records every attempt and its result.