What runs here
The server runs Qwen through Ollama, the NutsNews local AI API, telemetry collection, backups, and the observability portal. It has 16 logical CPUs and about 59 GiB of usable memory.
How it is managed
GitHub applies the recorded configuration to the existing machine. Normal releases do not wipe disks, reinstall the operating system, or delete models. The AI services are left running when changes concern only documentation or the portal.
What to watch
Look at CPU, available memory, storage, service state, and backup freshness together. A low CPU reading alone does not prove an application is healthy.