Updates endpoint for web container readiness check

[tannerjfco]

The Problem:

#66 original issue desc. We need to reliably determine if the web container is ready to serve.

The Fix:

This updates to check the new /healthcheck endpoint added to the nginx containers, rather than the root of the user's web application. This will allow us to determine container readiness regardless of the web app's state.

This also restores the Wait() in the start cmd, and updates restart to remove Config(), as that should have been removed.

The Test:

• Pick your favorite test site and run a a ddev start. Previously, the start cmd did not check readiness. You should now see that it is doing so.
• Run a ddev restart. Previously, this would have written a settings.php or wp-config.php, and would have failed status check as it was checking an uninitialized site. There should be no config file written, and the status check should succeed.

Automation Overview:

The local plugin tests have been adjusted to include Wait() in TestLocalStart() and TestLocalRemove().

Related Issue Link(s):

#66
drud/docker.nginx-php-fpm#33
drud/docker.nginx-php-fpm-local#7

Release/Deployment notes:

Does this affect anything else, or are there ramifications for other code? Does anything have to be done on deployment?

These PRs are part of this change:
drud/docker.nginx-php-fpm#33
drud/docker.nginx-php-fpm-local#7

Provisional tags for v0.3.1 have been pushed to dockerhub. Upon merge in both repos, a new v0.3.1 release should be tagged.

[rfay]

I tested this manually and am good with it, so approving. Please consider if you can or should add tests for this feature.

[rfay]

No biggie, but "couple of minutes" would be more idiomatic. The actual total time on my 2013 Macbook Air was 27 seconds total for ddev start from beginning to end, but the containers had already been cached. However, all that is before this statement. Might even be OK to say "a minute or more" here.