I have been looking into event-driven architecture for asynchronous communication between services and a design pattern I see quite often is the Single Bus*, Multi Account pattern, where services publish events to a Global message router EventBridge event bus, which then routes these events to other service's local EventBridge event buses based on Rules and they can decide how they want to use the event. A bunch of AWS Serverless talks mention the advantages but none have talked about how a central router bus like this can be a single point of failure. I am looking for ways I can minimize this risk by adding redundancy or otherwise.
AWS talks referenced: