The caddy, haproxy, openresty, and traefik docs stated that requests are only routed to containers passing healthchecks, but these proxies route to a new container as soon as it starts. The first `web` readiness healthcheck in `app.json` is now also turned into health checks and retries for caddy and openresty, as it already was for traefik, so requests are routed away from containers that are not yet ready, and the docs now describe how each proxy actually behaves.
Concurrent commands for the same app could each read a setting, change it, and write it back, so all of them exited successfully while silently dropping changes made by the others. Updates to app settings are now serialized so every change is kept, and readers never see a partially written value.
While I do not agree with _every_ style change, this will force Dokku to have consistent formatting across all shell scripts, which is arguably a Good Thing™.
The command used to reprocess everything is:
```shell
shfmt -l -bn -ci -i 2 -w .
```
In some cases, it may be useful to completely override the set ports for a given application - such as on initial deploy when there is no global domain set. This change also further allows us to handle automation in a smoother fashion.
Rather than removing all port mappings, we simply remap port 80 mappings to 443. The previous behavior would remove custom port mappings for all applications and reset them on any certificate change, preventing automation from plugins such as letsencrypt. While this behavior doesn't matter for buildpack deploys - which only expose a single port on container port 5000 - Dockerfile deploys would frequently be affected by such a change, requiring a remapping of all custom ports.
This commit also standardizes on the method used to source the proxy functions