Epic: drop vm-monitor in favour of neonvm-daemon

[Omrigan]

Motivation

A few from the top of my head:

• vm-monitor has some amount of tech debt (see e.g Epic: vm-monitor should check postgres during health checks #619, Epic: Centralized vm-monitor state #620, and other)
• Maintaining vm-monitor in neon repo, separate from autoscaling is a bit awkward (e.g. needs release coordination, separate CI, etc)
• Having vm-monitor in Rust, while the rest of the autoscaling in Go also makes it harder to maintain.

But we should further evaluate this.

DoD

vm-monitor no longer exists, it's functions are handled by neonvm-daemon.

Implementation ideas

TODO

Tasks

Tasks

Preview Give feedback

1. ...
2. List tasks as they're created for this Epic

Other related tasks, Epics, and links

• Slack thread: https://neondb.slack.com/archives/C03TN5G758R/p1727893026154979

[stradig]

We should start by writing an RFC for this, so that we can estimate the complexity.