Old CRT type oscilloscopes were nothing but analog devices. They were filled with potentiometers. All this circuitry would drift over time. When they were calibrated, known waveforms were fed to the scope, the displayed wave form was then compared to the known and if the differences was within a given margin the pots would be adjusted to bring it back to the original spec. This procedure depends on that the scope has a predictable drift. If the difference between the known waveform and the displayed waveform was significantly greater than the expected drift, something was wrong with the scope and it needs repaired.
New digital oscilloscopes don't have potentiometers like the old CRT scopes. The verification process, I assume, is exactly the same. Feed the scope known waveforms and compare them with the displayed waveforms.
With no potentimeters how is the scope adjusted for drift over time?
Is the scope trimmed in software or if the scope is out of tolerance (but still within the excepted drift) it needs repaired?