Check latency after QoS changes and a mesh restart

You changed QoS settings, restarted a mesh node, and now calls or games feel different. A single speed test may not explain why: the connection can be fast but develop long delays when someone downloads or uploads.

Use a bufferbloat test to check how much latency rises under load. Then compare the result with your previous readings, if you saved them, and check whether the mesh connection itself is stable.

Test latency after the restart

1. Open the Bufferbloat Test on the same computer you used before, if possible. Keep the computer in the same room and connected to the same mesh network.
2. Pause downloads, cloud backups, video streams, and other heavy network activity on your devices. Leave the mesh node in its usual location and confirm in your router or mesh app that it is online.
3. Run the test and note the idle latency, latency under download and upload load, and the overall grade. The grade runs from A+ to F. If you have older results, compare the figures rather than relying on the grade alone.
4. Run the test again after a short pause. If the results differ sharply, check whether another device started transferring data or your computer switched to a different mesh node.

For a fair comparison, change one thing at a time. If you are checking the effect of QoS, do not also move the mesh node, switch Wi-Fi bands, or change computers between runs. If you only started measuring after the restart, you can assess the current connection, but you cannot confirm that QoS improved it without a comparable earlier reading.

What the result means

The grade is good and latency barely rises under load

Your connection is handling the test's download and upload traffic without adding much delay. This is a good sign for calls and games while other devices use the internet. It does not prove that there will never be a stall, or that every room has equally strong Wi-Fi.

The grade is poor and latency rises during a download or upload

This points to bufferbloat: queues build up when the connection is busy, delaying other traffic. QoS, or quality of service, is a router feature that can prioritize or manage traffic. Check that the setting is still enabled after the restart and that any bandwidth limits reflect the speeds your internet connection actually reaches.

Idle latency is already high

A poor idle reading is not, by itself, evidence of bufferbloat. The delay may come from the Wi-Fi link, the mesh backhaul, your internet provider, or the test route. QoS is mainly meant to manage delay under load, so it may not fix a high baseline.

Results vary a lot between runs

A device may be roaming between mesh nodes, or background traffic may be changing. A node restart can briefly affect which access point a device uses. Check the connected node in your mesh app, then repeat the test from the same location when other network activity is quiet.

The bufferbloat grade looks fine, but you still get pauses

The test measures added delay during its workload, not every kind of Wi-Fi or internet fault. To look for repeated response-time spikes, jitter, lost requests, or dropouts over several minutes, run Connection Stability while you use the network normally.

How to fix it

1. Stop other large transfers for a clean retest. Pause cloud sync, game updates, and video uploads on all devices. Run the bufferbloat test again before changing router settings.
2. Confirm which network and mesh node you are using. On Windows 11, check Settings › Network & Internet › Wi-Fi to confirm the connected network. On macOS, check System Settings › Wi-Fi. Use your mesh app to see the connected node, if it provides that detail.
3. Check that QoS is still enabled on the router. QoS is configured in the router or mesh app, not in Windows or macOS. After a node restart, verify the setting on the primary router and confirm the service or bandwidth limits have not reverted.
4. Recheck the bandwidth limits. If the router shapes traffic using a download or upload limit, compare those values with recent real-world speeds. A limit set far above what the connection can deliver may not control the queue effectively. Change one value at a time and retest.
5. Compare Wi-Fi with a wired connection. If possible, connect the same computer to the main router with Ethernet and repeat the test. If the wired result is steady but Wi-Fi is not, focus on mesh placement, the node's link to the router, or client roaming rather than changing QoS again.
6. Check the mesh node's placement and status. Make sure it has finished restarting, is not showing a weak backhaul link in the mesh app, and has not been moved behind a cabinet or other obstruction. A node with a poor connection to the main router can pass that problem on to devices connected through it.
7. Restart and update in a controlled order. If settings are correct but results remain inconsistent, follow the mesh maker's restart order for the main router and nodes, then check for firmware updates. Avoid repeatedly restarting during measurements, since each restart can change the connection temporarily.

Why a mesh restart can change the result

A mesh node is one part of a chain: your device connects to a node, that node forwards traffic through the mesh, and the main router sends it to the internet. Restarting a node can change how devices attach or how traffic travels between nodes. A latency change after a restart therefore does not automatically mean the QoS setting changed or failed.

Bufferbloat and mesh coverage are related only in the sense that both affect how the connection feels. A bufferbloat grade describes added delay under the test's load. It does not show Wi-Fi signal strength in dBm or identify which mesh link is weak. If delays continue when the network is idle, investigate connection stability and mesh coverage separately.