When I try to monitor outlook.office.com:443 in netpath, it shows ~50% packet loss to all nodes. Use of the service shows that this is clearly not actually happening for user traffic. What's up with this?
Just speculating here based on how MS does a lot of azure services, but is it possible it's because of how outlook operates in that netpath isn't recognizing something? MS's cloud infra does modify packets for their internal networks. Tagging cobrien for this one.
I haven't seen this type of behavior before. That means it will probably require deeper investigation (diags, path export, potentially packet capture), so a ticket to Support is probably your best bet.
