Carrier's have commonly delivered 10BaseT and 100BaseTX Ethernet services with auto-negotiation disabled, instead manually configuring the speed and duplex on the handoff port.
This can be problematic if such a link is plugged into an RJ45 port with autoneg switched on. At the time it was standardized, auto-negotiation for these PHYs mandated that should the process fail, devices would default to half-duplex operation. Hubs were still common in that era and this was seen as a "safe" default. Unfortunately in today's world that means if a link has one side hard-set to 100Mb/full-duplex, and the other set to negotiate, the side set to autoneg will end up in half-duplex mode. This duplex mis-match means if the half-duplex side receives data while it is transmitting it will count that as a "collision", drop the frame it's recieving, stop transmitting, and start a CSMA/CD back-off timer. This kills performance.
Anyway enough showing my age. We have this issue on mr1-ulsfo, as can be seen from the device output:
cmooney@mr1-ulsfo> show interfaces ge-0/0/0 Physical interface: ge-0/0/0, Enabled, Physical link is Up Interface index: 134, SNMP ifIndex: 508 Link-level type: Ethernet, MTU: 1514, Link-mode: Half-duplex, Speed: 100mbps, BPDU Error: None, MAC-REWRITE Error: None, Loopback: Disabled, Source filtering: Disabled, Flow control: Enabled,
To properly terminate this link we need to configure our interface for 100Mb/full-duplex manually, disabling auto-negotiation for the port.
Agreed way forward is to set the Netbox interface type to "100BaseTX", indicating it needs to be set to that speed manually. Creating this task to track the homer and template changes to add this to our automated configuration.