Picture a shared meadow. Beautifully lush and green. Everyone lets their herd graze there, and for a while all's fine. Then the grass runs out, for everyone at the same time. So you fence the meadow into private plots, and now you've lost the reason you were sharing it in the first place.

Fabric capacities work the same way. It doesn't matter much whether your organization runs one big capacity or multiple smaller ones: a single runaway notebook or a heavy refresh is enough to slow down everyone's reports in that same capacity. Giving every team their own capacity doesn't solve that either. You lose the benefit of sharing, and the waste just shows up in more places. However you split it up, you still have to manage the whole meadow as one system.

In this session we look at capacities from a system perspective, and determine which feedback loops we can add to manage it better. We stay close to Fabric, so smoothing, bursting, throttling and surge protection all come back, just as parts of a system instead of a list of features. What you take home applies to any shared resource where the people using it can't see what it costs everyone else.

Then we get practical. Chargeback is the lever most organizations reach for first, and we'll discuss why the monthly, departmental kind is one of the weakest ones you have. Near-real-time feedback to the people actually causing the load works much better. Stronger still is something no tool gives you: the organization realizing it's sharing a resource. You'll leave with a diagnostic you can run on your own environment: which feedback loop is broken, and who is blind to their own impact?