URGENT: ECS Fargate Slaves channel closed exception randomly during pipeline execution

Hi Team,

we are using ECS Plugin and configured it to use Fargate slaves. While the pipeline jobs are getting executed the slave is getting disconnected randomly. We are seeing this error for 2 out of 5 jobs that are getting executed. We don't see any spike in Memory of the container when it's getting disconnected. As we are expecting to use the plugin in production we'd like to hear a permanent fix for this issue. Here are the details of our setup

Jenkins version: 2.440.1 ECS Plugin version: 1.49 Java Version: Java 11

Container slave image has latest JNLP base image
Jenkins Master running in EC2 instance behind ALB (us-west-2)
TCP fixed port 40272 is opened at both ALB and EC2
ECS slaves were getting created in both us-east-1 and us-west-2 regions as we have seperate plugin configs for each region which indeed proves no firewall issue

In ECS plugin configuration,

Tunnel through: :40272 Alternative jenkins URL: http://:8080/ (We also tried with https://jenkins-domain-url/)

Error:

FATAL: command execution failed hudson.remoting.ChannelClosedException: Channel "hudson.remoting.Channel@39b9f5b3:JNLP4-connect connection from x.x.x.x.us-west-2.compute.internal/x.x.x.x:57600": Remote call on JNLP4-connect connection from x.x.x.x.us-west-2.compute.internal/x.x.x.x:57600 failed. The channel is closing down or has closed down

Please find the attached screenshots with plugin configuration

jenkinsci / amazon-ecs-plugin

URGENT: ECS Fargate Slaves channel closed exception randomly during pipeline execution #348