VSZ, RSS, PSS, and What They Mean

Suppose a production dashboard shows these memory numbers for an API process:

So how much memory is the process actually using?

All three numbers can be correct because they measure different things:

Memory analysis becomes confusing when these metrics are treated as interchangeable. A large virtual address space does not necessarily mean high RAM usage, and a heap dump does not explain the process’s entire memory footprint.

This chapter builds a practical model of the most important process memory metrics: VSZ, RSS, PSS, and USS, and shows when to use each one.