Contradictory sizing requirements for control plane nodes

[eduardolat]

Bug Description

The documentation in /docs/installation/requirements.md contains contradictory sizing information for control plane nodes that confuses users.

Contradictory Information

Server Sizing Guide table states:

• 2 CPU, 4 GB control plane → supports 0-350 agents

Large Clusters - CPU and Memory table states:

• Small clusters (up to 10 nodes) → 2 vCPUs, 4 GB per control plane
• Medium clusters (up to 100 nodes) → 4 vCPUs, 8 GB per control plane

The Problem

These tables contradict each other:

• First table: 2 CPU, 4GB can handle 350 workers (50% more in HA)
• Second table: 2 CPU, 4GB only good for 10 total nodes

It's unclear whether "nodes" in the second table includes control planes or refers only to total cluster size.

Suggested Fix

Clarify the documentation by:

1. Defining whether "nodes" includes control planes or workers only
2. Explaining why requirements differ between the tables
3. Providing consistent terminology throughout

Impact

This confusion leads users to incorrectly size their control plane nodes and creates uncertainty about k3s scaling capabilities.

[dereknola]

The large clusters information needs to be combined/eliminated as its fairly old information. The First Table is the more up to date info.