fix healthcheck in dockerfile

[thrau]

Motivation

As a user pointed out in #10893 (comment), the healthcheck command was faulty after #10800:

thomas@ninox /tmp/foo % docker exec -it foo-localstack-1 bash
root@cb3bded7f1f7:/opt/code/localstack# ./bin/localstack status services 
Traceback (most recent call last):
  File "/opt/code/localstack/./bin/localstack", line 23, in <module>
    main()
  File "/opt/code/localstack/./bin/localstack", line 17, in main
    from localstack.cli import main
ModuleNotFoundError: No module named 'localstack'

I'm deliberately not fixing bin/localstack because I believe we should not be using it directly in this way.

I also noticed that rolo is logging HTTP health check requests again, so I updated the default log levels to what we had in localstack.aws.serving.

Changes

• HEALTHCHECK in Dockerfile now uses localstack from the .venv instead the plain script, which uses the correct pythonpath
• default log levels for rolo is now WARNING and DEBUG only for tracing

[dominikschubert]

LGTM, tested locally as well to make sure it works. Was confused at first why that should work but looking at the contents of the script it makes sense 🙃