Experiencing the same problem on two CRS354-48G-4S+2Q+ devices, tried multiple routeros / firmware versions (6.46.4, 6.46.5, 6.46.6, 6.47 and 6.47.1), now trying 6.48beta12..
disabling -> enabling (individual) ports does not seem to resolve the issue, rebooting the whole device does resolve the issue (temporary).
sometimes it magically get resolved automatically after 'a few minutes ~ a few hours' but most of the time is does not auto-recover.
In the beginning we've experienced the issue 'once a week' (after a few days), now we experience the problem like every '± 4 hours'... Not sure if this is related to 'more devices connected', 'more throughput' etc.
I also disabled spanning-tree (RSTP) and 'Fast Forward
' (keeping Fast-Path / hardware offload
on) on the bridge, and loop-protection on the individual ports but this is not the problem / not resolving the issue.
Interesting thing is that the log does not show anything related to the issue.
In this thread I've read that it seems to only affect ports 1-8, it seems that we can confirm this!
The switches are mounted in the rear of datacenter rack, regarding health / temperature:
But I'm not sure if it's cooling / temperature related since the other ports (9 - 48 and SFP / QSFP) are working fine, also management is working fine and quick / responsive.