How can RabbitMQ's (Pika) unack redelivery time be increased? We have a long running task that takes longer than the default redelivery time, and so when it tries to ack once the processing is finished, it's causing an unable to send ack error.
I'm assuming you're running into the channel ack timeout that is discussed in full here - https://github.com/rabbitmq/rabbitmq-server/pull/2990#issuecomment-1002089576
Long-running tasks like yours should publish the message being worked on to another queue representing in-progress work (perhaps a queue per worker), then ack the original message. When the work completes, consume and ack the message in the in-progress queue, and re-consume from the original queue.
Yes it is more work but the channel ack timeout was introduced for a very good reason. If you know you won't run into the issue described, you can increase the timeout in RabbitMQ's configuration.