Dynamic port mapping - health check failures for target group in ECS cluster

We have our application hosted in ECS cluster as container. Also we have nginx running in a container to reverse proxy. 0.0.0.0:80->8080/tcp - application 0.0.0.0:81->80/tcp - nginx

I have a target group having my instances running on port 80 and health check settings configured to use override option and port mentioned as 81 - so that health passes

I am trying to use ECS blue green code deploy option in code pipeline to automate deployments. To achieve this, I have changed my application port mapping to dynamic so that it runs in a random port - for ex: 0.0.0.0:32768:8080

I created a task definition to use dynamic port mapping like 0:8080

I have an application load balancer which listens on 80 and 443 and redirect traffic to target group .

While trying to create a service using the above ALB andd TG, I am getting error as the task definition uses dynamic port mapping but the target group uses particular port specified in health check settings.

So I have changed the health check setting option to use traffic port option - but doing so the health check always fails.

Could you please help me in configuring a health check for dynamic port mapping or how to configure the resources for blue green deployments to my container.

Note : My application does not need any task to be running and its just deploying once. Verified security group inbound rules to allow traffic from load balancer