Giant Single Stack

One huge config that tries to run everything for everyone usually degrades developer throughput.

Typical symptoms

• startup is slow even for small tasks
• laptops run hot under normal iteration
• logs are too noisy to isolate incidents quickly
• teams restart each other's dependencies unintentionally

Why it fails

• dependency graph grows denser than needed
• optional services become mandatory cost
• local feedback loop is dominated by orchestration overhead

Better approach

• keep a core stack for daily work
• split optional or role-specific services into additional configs
• document which workflow uses which config

This is a reliability and speed issue, not just preference.

See When to Split Stacks.