I am running a site-to-site setup with Zerotier across two Synology NAS boxes. I have followed the instructions for Synology here: Synology | ZeroTier Documentation on both boxes, and then set up routing as indicated here Route between ZeroTier and Physical Networks | ZeroTier Documentation on both machines as well.
The connection itself works great, and is perfectly stable. Both subnets can reach each other without issues. However, I am running into severe issues with packets larger than ~1150B. For those packets, the connection is highly unstable and will work for a minute before dropping out for multiple minutes. During this dropout, I can still send/receive smaller packets. As a test, I am literally running two ping commands simultaneously to the same host - one ping a.b.c.d and the other ping a.b.c.d -s 1500. The small packets will never have any issues, while the larger packets will randomly stop being received at all for minutes at a time, before resuming again. These problems are not ICMP-specific - the same behaviour can be observed for UDP and TCP streams. Very occassionally, the small packets will also have large ping delays - from 18ms they jump to over 200 for a few seconds.
I have already tried setting a lower MTU, to 1280 in the API. I have confirmed that this is applied to the zt interface through ip a. I tried reinstalling the Docker container and rebooting the NAS boxes, to no avail.
Is anyone here able to help me out?