Disable in-container health checks for ad-hoc containers

[mik-laj]

This slows things down a bit if we're trying to write scripts that can sometimes run multiple containers. The health of the containers is already checked by Docker, so we don't have to check it on our side every time.

Comparision:

$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=0 airflow-worker airflow celery
0.66s user 0.14s system 23% cpu 3.396 total
$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=0 airflow-worker airflow celery
0.66s user 0.14s system 22% cpu 3.576 total
$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=0 airflow-worker airflow celery
0.61s user 0.13s system 22% cpu 3.210 total

$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=1 airflow-worker airflow celery
0.67s user 0.18s system 19% cpu 4.456 total
$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=1 airflow-worker airflow celery
0.63s user 0.14s system 18% cpu 4.182 total
$ time docker-compose run --rm -e CONNECTION_CHECK_MAX_COUNT=1 airflow-worker airflow celery
0.65s user 0.15s system 18% cpu 4.373 total

Part of: #11161

---

^ Add meaningful description above

Read the Pull Request Guidelines for more information.
In case of fundamental code change, Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in UPDATING.md.

[github-actions]

The PR most likely needs to run full matrix of tests because it modifies parts of the core of Airflow. However, committers might decide to merge it quickly and take the risk. If they don't merge it quickly - please rebase it to the latest master at your convenience, or amend the last commit of the PR, and push it with --force-with-lease.