fix(core): alert user to missing global services

[SamJUK]

Check List

• Is there an existing issue that covers this? (link it here, if so)
• Is there an entry in the CHANGELOG.md file?

Describe the bug
If a core service is missing, you get a generic no such container error message, and the application will still attempt to run.

To Reproduce
Steps to reproduce the behavior:

1. Checkout latest release 0.15.0
2. Start warden warden svc up
3. Checkout latest main branch
4. Start a project warden env up

Note: If you have previously been using latest dev branch, you may need to fully remove the phpmyadmin container.
docker rm -f $(docker ps -a -q -f name=phpmyadmin)

Expected behavior
Fatal error to occur to indicate to the user, that the core is in a unhealthy/outdated state.

Screenshots
Before:

After:

[bap14]

Given that PMA can be disabled using WARDEN_PHPMYADMIN_ENABLE=0, I would not consider it a "core" service to make the system work. It's an add-on that can be opted out of. PMA will not prevent the environment from running, or rather it shouldn't.

It would rather be better to refactor the peered service(s) to be only those that can't be opted out of (e.g. traefik, mailpit), and a separate process for "optional" services (e.g. portainer, dnsmasq, PMA) that first checks if they're enabled, and if so would add them to peered services list.

[SamJUK]

I wasn't aware of the feature flag, thanks for bringing that up. I agree the peered services implementation could do with some improvements, to better support optionally attached services.

I'll open a separate issue/PR to address that, and keep conversations focused.
#892

Although I still think the alert is a valuable addition, dependent on resolving the optional services issue. As it is essentially saying the global services state is not as expected. EG:

• A Global container has been killed by OOMKiller / stopped outside of warden
• New mandatory core service has been added by not yet started etc