apache – When kafka node goes down from cluster facing issue with kafka streams connected this kafka cluster

I am using 3 nodes kafka cluster, And there are multiple kafka streams and producer’s are connected.
I am facing issue when one of the node is going down from this kafka cluster.

Issue i am facing is like multiple kafka streams are in connected state but they are not consuimg messages when one of the kafka node’s goes down. All messages are in kafka topic queue even streams are in connected state.

Details :-

Each topic is configured with replication_factor as 3.
Each topic is having 6 or more than 6 partitions.
All 3 kafka nodes are configured as Broker and Controller.
Number of kafka stream thread’s are 3 and more.
Kafka cluster is deployed in kubernetes envi.
In bootstrap server mentioned all kafka node’s address as well.

I have added below properties in kafka stream and done testing by deleteing kafka node pods.

Properties :-

(StreamsConfig.REQUEST_TIMEOUT_MS_CONFIG, 30000).
(StreamsConfig.RETRY_BACKOFF_MS_CONFIG, 3000).
(StreamsConfig.TASK_TIMEOUT_MS_CONFIG, 300000).

But still facing same issue, That kafka streams are connected but not consuimg messages and message’s are stored in kafka topic queue.

To start consuimg messages it is required to restart the kafka stream.

Can some one help what changes can be made, So that kafka stream’s keep consuming messages even one of the kafka node goes down.