Rabbitmq sucks to scale. clustering and partitioning were terrible for a long time, maybe still is. Clusters dying in split brained ways, nodes crashing terribl

Rabbitmq sucks to scale. clustering and partitioning were terrible for a long time, maybe still is. Clusters dying in split brained ways, nodes crashing terribly / unrecoverable if they exceed iops or storage limits. You couldn’t pay me enough to run a high volume rMQ cluster again. Never mind that the persistent durable log pattern of kakfa enables a lot replay type use cases that are very beneficial in financial systems specifically. It’s not solely because of hype at all, it’s objectively better for many use cases.