I have deployed my RabbitMQ cluster with two replicas in my EKS. One replica runs on each node. As part of the Kubernetes version upgrade, I need to upgrade the nodes on which RabbitMQ is running. However, this causes the RabbitMQ cluster to go down. Is there a way to update EKS nodes without disrupting the RabbitMQ service?
I would suggest combining Pod Disruption Budgets PDBs for your RabbitMQ pods which will ensure at least your two RabbitMQ pods arealways running for example and the managed nodes rolling update strategy: