Compare Ethernet with Wi-Fi
Connect the affected device by Ethernet for one test.
If the spikes stop: fix the Wi-Fi path.
Why do I have lag spikes? Sudden ping jumps and rubber-banding usually come from Wi‑Fi instability, busy router queues, packet loss, household traffic or external routing problems. Compare the timing and network layer before changing hardware.
Lag spike troubleshooting guide
Lag spikes are brief jumps in latency. Use this guide to identify whether the sudden ping increase comes from Wi-Fi, household traffic, router queues, packet loss, your provider or one game server.
Lag spike meaning
A lag spike is a short, sudden rise in latency that makes a game, call or live connection pause, rubber-band or briefly disconnect even when the average broadband speed looks healthy.
Common causes include unstable Wi-Fi, router queues filling during downloads or uploads, brief packet loss, household traffic, peak-time congestion and changes in the route to the game or service.
Start here
Lag spikes are brief, intermittent jumps in latency: ping suddenly rises and falls, games rubber-band or freeze, and voice chat may break up at the same moment. A fast speed test does not rule them out. The useful question is what changes when the spike appears: connection type, household load, time of day or the game being used.
Constant rather than intermittent? If latency stays high instead of briefly jumping, use the high ping troubleshooting guide.
Connect the affected device by Ethernet for one test.
If the spikes stop: fix the Wi-Fi path.
Stop downloads, uploads, cloud backups, CCTV uploads and streaming.
If the spikes stop: household load or bufferbloat is responsible.
Test another game, voice app or server region at the same time.
If only one service is affected: investigate that server or route.
Guide index
Identify and fix
The symptom, likely cause, first useful test and matched fix are combined here so the page does not repeat the same advice in a separate fix section.
| What you notice | What it usually means | Best next test | Matched fix |
|---|---|---|---|
| Ping jumps, rubber-banding or voice breakup stops on Ethernet | Wi-Fi instability. Weak signal, interference, crowded channels, mesh roaming or device behaviour is forcing retransmissions. | Keep Ethernet as the control, then retest the same device over Wi-Fi in the same time window. | Keep latency-sensitive devices wired where practical. Otherwise improve router position, reduce mesh hops, avoid weak extenders and work through the unstable Wi-Fi fixes. |
| Spikes begin when another device uploads, downloads, streams or syncs | Bufferbloat or household contention. Gaming and call traffic is waiting behind bulk data, often because the smaller upload capacity is saturated. | Run the loaded-latency test with the connection quiet and busy. | Restart one activity at a time to find the trigger. Schedule large transfers outside gaming or call time and limit upload-heavy devices. |
| Idle ping is stable but loaded latency is high | Router or access queueing. The line may be fast when quiet but real-time packets wait when downloads or uploads fill the queue. | Repeat the bufferbloat test after each router or traffic-control change. | Enable SQM with Cake or FQ-CoDel where supported. Shape slightly below repeatable wired throughput, then follow the bufferbloat and SQM guide. |
| Rubber-banding appears with packet loss, voice breakup or short disconnects | Packet loss or unstable hardware. Poor Wi-Fi, damaged cabling, powerline adapters, router faults or the external line may be dropping data. | Run the packet-loss test over both Wi-Fi and Ethernet. | Replace the test cable and remove weak links. Bypass powerline adapters, docks, extenders and extra switches, then test directly over Ethernet. |
| Only one game, server region or voice service is affected | Destination server or route. Server load, distance, matchmaking region, VPN routing or one provider path can fail while the rest of the connection remains stable. | Compare another real-time service, select a nearer region and test without a VPN. | Do not replace the router yet. Check service status, change region, disable VPN routing and compare a second game before changing local hardware. |
| Wired spikes repeatedly become worse at a similar evening time | Peak-time load. Household usage, provider backhaul or an external peering route may be busy; one evening result is not enough. | Pause household traffic and compare the same wired test off peak and during the problem period. | Collect comparable quiet-time and peak-time evidence. Record idle, average and maximum ping, jitter, packet loss and affected services before escalating. |
| Spikes follow one device or repeat at a predictable interval | A device or background process. Cloud sync, updates, wireless power saving, drivers or periodic network scans may briefly interrupt traffic. | Test the device on Ethernet and close background utilities before changing the router or broadband package. | Stop the local trigger first. Pause cloud backup, launcher downloads, updates or power-saving features, then retest the same service. |
Use the matched fix only after the test confirms the layer. Changing DNS, replacing the router or upgrading broadband speed before the pattern is proven can hide the real cause.
Validate
Follow the steps in order and change one variable at a time. Record the result before moving on.
Use these as practical wired reference bands when testing a nearby UK server. Distance to the server changes normal ping, so the most important signal is a repeatable jump above your own quiet-network baseline.
| Connection pattern | Ping and spikes | Jitter | Packet loss |
|---|---|---|---|
| Very stable | Usually below 30 ms to a nearby server, with only small changes from the baseline | Below 5 ms | 0% |
| Needs investigation | 30–80 ms may be usable, but repeated jumps of 30–100 ms above the baseline can cause noticeable stutter | 5–15 ms | Any repeatable loss below 1% should still be investigated |
| Severe instability | Repeated jumps above 100 ms from the baseline, especially when maximum ping reaches 150 ms or more | Above 15 ms | 1% or more, or short bursts of consecutive loss |
Do not judge the line from average ping alone. A 20 ms average can hide repeated 200 ms peaks. Record idle, average and maximum ping together.
Record a quiet-network baseline. Run the LinkSpeed Ping Test while no one is downloading, uploading or streaming. Note idle ping, jitter and any packet loss.
Repeat the test while the spike is happening. A large change in maximum ping or jitter confirms an intermittent stability problem even when download speed remains high.
Compare Ethernet with Wi-Fi. If Ethernet removes the spikes, focus on router position, channels, mesh placement, drivers or the affected device. The broadband line is not the first suspect.
Pause household traffic and retest. Stop console downloads, cloud backups, phone photo sync, CCTV uploads, livestreaming and large file transfers. Immediate improvement proves local contention.
Run the Bufferbloat Test. If ping rises sharply during upload or download load, the router or connection is queueing traffic badly.
Check for packet loss. Use the Packet Loss Test over Ethernet and Wi-Fi. Loss only on Wi-Fi points to the local radio path; wired loss across services needs wider investigation.
Compare another game, server region or real-time app. If only one destination is affected, do not replace the router or broadband package before checking the service and route.
Compare peak and off-peak results. Keep the same device, cable and test destination. Repeated evening-only wired degradation is stronger evidence of provider-side congestion than one isolated result.
Advanced verification
The LinkSpeed Ping Test measures repeated HTTPS request times to LinkSpeed. A native ping command normally sends ICMP echo requests to the endpoint you choose. The results are not directly interchangeable, but using both helps show whether the spike follows the browser, the home network or the external broadband path.
Open two command windows and run both tests for 60–120 seconds while the lag is happening. Replace YOUR_ROUTER_IP with the router address shown by ipconfig in Windows, Network Settings in macOS or ip route in Linux. Common examples are 192.168.0.1 and 192.168.1.1.
Windows Command Prompt
ipconfig
ping -t YOUR_ROUTER_IP
ping -t 1.1.1.1
macOS or Linux Terminal
ping YOUR_ROUTER_IP
ping 1.1.1.1
Press Ctrl+C to stop. Compare packet loss and the minimum, average and maximum delay; the exact summary fields vary by operating system.
Run a route trace after repeatable external spikes have been confirmed:
Windows
tracert 1.1.1.1
macOS or Linux
traceroute 1.1.1.1
An asterisk or one slow intermediate hop does not by itself prove a fault because routers can ignore or deprioritise diagnostic replies. Stronger evidence is delay or loss that begins at one point and remains visible through later hops and the destination.
Need a simpler first pass? Use the LinkSpeed Ping Test to capture a repeatable latency graph, then use the command-line comparison only to verify the fault boundary. For longer failed-request sampling, run the LinkSpeed Packet Loss Test.
Open Task Manager and sort processes by network use while the spike happens. Pause OneDrive or Dropbox sync, Windows Update and Delivery Optimisation, game-launcher downloads, browser uploads and cloud-backup software, then retest.
Compare the console's network statistics over Wi-Fi and Ethernet and note packet loss, connection speed and NAT status. NAT type can explain matchmaking or voice-chat problems, but it does not by itself prove the cause of intermittent ping spikes.
A packet-burst, packet-loss or latency-warning icon that appears with rising ping or loss points to the network. If ping stays flat while the game freezes or frame time rises, investigate CPU, GPU, storage, drivers or thermal throttling instead.
Using 5G broadband? Radio signal, mast load and router placement need a different validation path. Use the 5G broadband lag guide.
DNS and static IP settings: changing DNS can affect name lookup and a static IP can simplify port rules, but neither normally fixes Wi-Fi retransmissions, bufferbloat, packet loss or provider congestion.
Escalate
Contact the provider when the problem remains on Ethernet, affects more than one game or real-time service, continues with household traffic paused and appears in repeated tests.
Provide:
A single in-game screenshot is weak evidence. A repeatable wired pattern across several services gives the provider a clearer fault boundary.
Replace the square-bracketed details with your results, then paste the message into provider chat, email or a fault ticket.
FAQs
A lag spike is a short, sudden rise in latency that can make a game, call or live connection pause, rubber-band or briefly disconnect even when average broadband speed looks healthy.
Download speed measures capacity, not consistency. Wi-Fi retries, busy router queues, upload saturation, packet loss or routing changes can delay a few time-sensitive packets without reducing the headline speed much.
The connection or router is probably queueing real-time traffic behind bulk data. Run the bufferbloat test and reduce background uploads and downloads before changing package.
Interference, weak signal, crowded channels, mesh roaming or device behaviour can cause brief retransmissions. A stable Ethernet result confirms the local wireless path is the better place to start.
Evening spikes can come from household activity, provider congestion or busy routes. Compare wired quiet-time and peak-time tests with other traffic paused.
Escalate when repeatable wired spikes affect several services with background traffic paused. Include times, ping, jitter, packet loss and the affected applications.
On a wired test to a nearby server, below 5 ms is very stable. Repeated jitter above 15 ms is likely to be noticeable, especially when maximum ping and packet loss rise at the same time.
It usually means game data arrived late, in a burst or not at all. Confirm it by checking whether ping, jitter or packet loss rises when the icon appears; the symbol alone cannot identify whether the cause is Wi-Fi, router queues or the external route.
Usually not. DNS mainly affects name lookup and a static IP affects local addressing. Neither normally fixes interference, bufferbloat, packet loss or peak-time congestion.
No. The LinkSpeed browser test measures repeated HTTPS requests to LinkSpeed, while Command Prompt or Terminal ping normally measures ICMP replies from a chosen endpoint. Use both because they test different traffic and destinations.