LAG Connections link speed incorrect on path

[Fuhunter]

Description

When running a LAN speed test, the LAG connection on one switch shows an incorrect link speed in the path overview.

Instead of displaying the aggregated bandwidth (2×10G), it only shows 2.5G or the slowest link speed. This appears to misrepresent the actual uplink capacity of the LAG.

Environment

• NetworkOptimizer version: 1.10.3
• Controller / firmware version: 7.2.123
• Switch model: USW-Aggregation and USW-Pro-Max-16
• LAG configuration: 2 × 10G ports (Aggregation)

Steps to Reproduce

1. Configure a LAG between two switches using 2 × 10G ports.
2. Start a LAN Speed Test between two clients across that LAG. Or WAN Speed Test from Server
3. See the path view after the test.
4. Observe the reported link speed for the LAG connection.

Expected Behavior

The path should display the aggregated LAG bandwidth (20G or clearly indicate 2 × 10G).

Actual Behavior

The path shows 2.5G as the link speed instead of the aggregated 20G.

Additional Notes

• Traffic flow itself appears normal.
• The issue seems to be limited to the displayed link speed in the path visualization.
• Screenshot attached: link between core aggregation and core switch

[tvancott42]

Nice catch, I think I have enough data from Security Audit knowing how to handle LAGs now that I can get this working. I'm stoked you actually caught this!

[Fuhunter]

No problem, just created this today :) In the UniFi UI there is a Group Index for Aggregated Ports. Maybe you can use that?
What I don't get is, why does it show the slowest link speed and not at least the proper link speed of that port?

[tvancott42]

That's a good question. I have the logic now sorted out, short of having a LAG on one of my own trunks, but it's difficult to say why 2.5 GbE was picked up instead of the base 10 GbE. If you turn debug logging on, there may be a clue.

The "fix" (I hope it works at least, it may in theory) will be included on the next release, probably within an hour or so.

edit: (disregard auto-close issue, that's from me merging the hopeful fix)

[tvancott42]

@Fuhunter this may be fixed now in v1.10.4: https://github.com/Ozark-Connect/NetworkOptimizer/releases

If not, please enable debug logging (https://github.com/Ozark-Connect/NetworkOptimizer/blob/main/docker/DEPLOYMENT.md#logging-configuration) and check the logs to see if there's anything useful about that link we can see.

[Fuhunter]

Hey,

I just tested it with your latest version, it is not fixed. I tested it on my Mac and tried turning on Debug logs but it did not work. Set log_level and app_log_level to debug in start.sh. Is there another place where I should set those variables?

[tvancott42]

@Fuhunter thanks for checking, I had a feeling this would be something beyond what my fix entailed. Here's how you enable debug logging on the native macOS installation: https://github.com/Ozark-Connect/NetworkOptimizer/blob/main/docs/MACOS-INSTALLATION.md#enable-debug-logging