The recent cloud services outage was triggered by issues the WMCS network storage (ceph) clusters had, caused by a lowered MTU on a L3 switch IRB interface.
While the particular issue exhibited signs of a bug rather than exactly what Juniper say ought to happen, it's clear there is a risk that if an MTU is not set on a L2 interface in Netbox, the switch port will not have one assigned, defaulting to 1514 and potentially causing issues with routing of jumbo frames.
Medium-term we should accelerate the work on T310590, to enforce a constraint that MTUs have to be set (or even set to a particular value), for L2 access/trunk ports. In the meantime it may be worth adding some other mitigations to prevent a mis-configuration similar to the one that caused the cloud outage.
Creating this task to track options and changes towards that end.