@Ethan-TP 

1. When did you first notice this issue? For example, did it exist before you upgraded the Gateway firmware?

  This has been flaky since the first 1.3.0 beta , but 1.3.2 official and the 1.3.2 august beta (with SNAT fix) is especially bad.

2. Please verify whether the problem is limited to this single port only, or if other ports exhibit the same behavior.

  I only have this one downlink to my main switch

3. Kindly provide the model of the device connected to this port and the exact SFP+ module part number you are using, so we can better analyze the issue.

  Generic 10g DAC cable going to SG3428X

4. Have you tried swapping the SFP+ module for a different model, or connecting this SFP+ port to a different switch (or to another port on the same switch)? If possible, please perform these tests.

  Yes tried several cables and different switch ports, even tried some TPLink SFP+ to RJ45 modules, same issue

5. You mentioned that “Omada control data is unaffected.” Please use ping tests to confirm that, apart from Internet traffic, all other internal-network communication (e.g., pings between devices) still works correctly.

  I can ping, i can continue general configuration in controller that effect thae gateway, i can access its web GUI.  All NATed internet traffic just stops. everything else is fine.

All my vlans are switch routed, so the gateway has a static route to a switch SVI on the SG348X, and when the issue occurs its the same effect as if the gateway static route is disabled or not present - no internet traffic.  Local traffic remains working.  However, im not disabling or editng the switch route, it has been the same for about a year, has a normal name (no special characters or spaces etc).