Your organization's plan grants a fixed compute allowance, and the dashboard shows exactly how much of it is in use - overall, and broken down per app instance.
Tracked metrics include compute (processor and memory), network traffic, and storage, both for your organization as a whole and for each instance individually. That per-instance breakdown is what lets you answer "which of my apps is actually using the most" rather than just seeing a single combined number, which matters once you have more than a couple of instances running.
See Concepts › Usage and quotas for how this ties to your plan, and Guides › Monitoring usage for where to look in the dashboard.