The problem is just half an hour back resolved. When a NLB is spun up by the application it use the combination of the namespace and a string to create target group to which it will forward the request.

There was ghost of older target group that existed in the AWS that cause duplicate names for the target group and thus the new NLB was not able to register the target group and thus create a listener.

Once we deleted this ghost target group from probably an old install, the NLB was able to register the target group and listener and the external IP  resolved to the FQDN of the NLB.

Post that it was usual of registering the Network interface to the ALB and accessing the application via ALB which worked smoothly.

I took an AWS expert to figure out that the error was due to duplicate target group that existed.

So one possibility is to have a check done by the application if these ghost of NLB and Target groups exist since they use same string names for the naming these target groups and if it does then inform to either clean up during check process or get clean by themselves using AWS commands. This would be suggestion to incorporate in future releases.