Managed platforms abstract the server away, until something breaks that the dashboard can't show. Knowing how to get on the box and look around is a skill worth keeping.
The basics cover most incidents: check the logs, check disk space, check what's listening, check what changed. `journalctl`, `df`, `ss`, and a recent diff answer more questions than most dashboards.
This isn't anti-platform, it's pro-fluency. The abstraction is convenient until it isn't, and the developer who can drop to the metal is the one who fixes the thing the dashboard can't see.
