Remove run.sh entrypoint for Prometheus

[reversesigh]

When running multiple DVnodes, it can be easy to integrate them into a single prometheus job like this:

  - job_name: charon
    static_configs:
      - targets: [charon:20120,charon:20220]
    static_configs:
      - targets: [lodestar:20164,lodestar:20264]
  - job_name: lodestar

However with the entrypoint for run.sh, these changes get overwritten whenever the container relaunches. Lido already has this removed from their fork by default - and I think it makes sense to make it easier to make adjustments to prometheus.yml by default.

[aly-obol]

@reversesigh This repo mainly deploys and monitors a single node. For multiple-node deployment, there is another repo where we don't use the Prometheus run script.

Also, this script centralizes all the config in a single .env file without manually editing the Prometheus yaml file to add the monitoring token and endpoint. We are open to changing the deployment while the goals above remain intact.