It's more to the point that disruptions tend to be very disruptive, too frequent, not well understood, and the reset takes a really long time. But you're right that this does not quite belong. If we solve the technological problems, the maintenance should resolve itself.
Making the reactor sufficiently maintainable is a technology problem.
There was a group that tried to analyze the expected uptime of DEMO, the pre-commercial follow on to ITER. With the expected MTBF (mean time between failure) and MTTR (mean time to repair) they estimated it would be working just 4% of the time. That's too low even for a research machine.
RAMI (reliability/availability/maintainability/inspectability) is one of the major issues for fusion, and even if nothing else stops fusion it alone will make the development process very long and iterative. Good RAMI requires long iterative improvement of individual components and processes.
