Reuse existing container image

A simple solution is using a configmap to alter such xyz.sh like:

apiVersion: v1
kind: ConfigMap
metadata:
  name: new-entrypoint
data:
  new_entrypoint.sh: |
    #!/bin/bash
    sed -i 's/Host:xyz\.com/Host:abc.io/' /path/to/xyz.sh
    exec <your-original-entrypoint.sh> "$@"

Then mount this new configmap and use the file new_entrypoint.sh in command attribute.

This is ugly but you don't need to maintain 1k+ lines of bash script into a configmap.

Notice: this may not work if your xyz.sh doesn't belong to current user/group that runs your container. Also it violate the idea of immutable infrastructure. So you will need to replace the xyn.sh via a configmap entry.

From scratch

So the line Host:xyz.com becomes configuration that you want to change at runtime. so you may want to alter such hard coded value into environment variable so that it can be override at runtime.

E.g., in your xyz.sh:

MY_HOST=${MY_HOST:-"xyz.com"}

In dockerfile:

ENV MY_HOST="default_host_for_your_docker_image".

In k8s pod description, you can do:

containers:
  - name: xyz
    image: "your-docker-image-name"
    env:
    - name: MY_HOST
      value: abc.io

See section III. Config of the 12-factor app:

Store config in the environment