Replacing bespoke scaler with keda; ScaledJobs that process multiple long-running messages

[charlesritchea]

I'm considering keda as a replacement for a bespoke python service that watches rabbitmq queues and submits k8s Jobs that consume rabbit messages until an idle timeout, e.g. 5minutes. There is considerable overhead in scaling a Job pod, so the recommended pattern of 1:1 message to Job is not ideal for us, and ScaledObjects don't work if they can just terminate the pod before work is complete. Is this possible now with Keda? The docs emphasize a single message