I have a function app with ServiceBusTriggers, it consumes messages from Service Bus at high throughput.
The function app scales automatically, so Azure might increase the number of consumers based on the messages on the bus.
In the scenario where a dependency (i.e the database) is down, I'd like the function app to back off and stop taking messages off the subscription entirely (else it'll throw thousands of messages on the DLQ)
But how can I achieve this when the function is auto-scaled? If I put waiting retry logic into the function app, Azure might think messages are processing slowly, and spin up more instances