How to distribute the load of subsystems?

Question

Software design

Hello!

Interested in the issue of managing sub-systems and regulating the load between servers.
Let me take a simple example: A service is a telegram bot constructor that hosts a bot for it on some server when the client presses the "START" button (or everything happens on the same machine with resource allocation).
When there are a lot of such users, then we have a large number of systems that operate exclusively in the space of each user and are able to communicate with the central node to process commands (Start, stop, etc.).

Actually, are there any relevant articles that propose an architecture for managing all these processes?

The simplest thing that comes to mind is a central node to which all available hosts are authorized. Each host has its own central control program (process). When requesting the launch of a new bot, the central node polls the load of free nodes and sends a command to one or less loaded nodes to start. The node, in turn, launches a docker container on its local machine and then manages it.

Answer 1

No need to reinvent anything - everything has long been invented in k8s and the CNCF community