Question ongoing ping spikes and packet loss

Aug 1, 2024
4
0
10
Hello,
as the title states I have been having very bad ping spikes and packet loss for the past 2 - 3 weeks and I can not seem to troubleshoot it. I have restarted my modem multiple times, I tried to factory rest the modem and disconnected all the devices that was connected to the wifi but none of the solutions worked. I've tried pinging www.google.com in command prompt and I get random ping spikes about every 30 seconds to 1 minute. the ping spikes can range from 120ms - 2200ms but exceeding 1000ms is rare. Ping plotter shows crazy packet loss and I'm not sure what to do about it.


network:
Type: Rogers 1.5 Gigabit Ignite Internet
Modem: Rogers XB7
Router: TP-Link Archer AX80
Connection: Cat 6 Ethernet connected to PC
Speed: 939.25 Mbps download 41.07 Mbps upload


PC:

Motherboard: ROG Strix B650-a
CPU: Ryzen 7 7800X3D
Ram: DDR5 6000
GPU: Nvidea 4070 ti


PingPlotter Results to www.google.com:
View: https://imgur.com/a/7L00wdR


from "ping www.google.com -n 100":
Ping statistics for 172.217.1.4:
Packets: Sent = 100, Received = 100, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 13ms, Maximum = 1570ms, Average = 40ms

from "tracert www.google.com":
1 * <1 ms * 1xx.xxx.x.x
2 * * * Request timed out.
3 24 ms 16 ms 24 ms 8021-dgw02.nmkt.rmgt.net.rogers.com [2x.xxx.xxx.xxx]
4 20 ms 24 ms 16 ms 100-0-18-0-5-cgw01.ym.rmgt.net.rogers.com [2x.xxx.x.xx]
5 16 ms 18 ms 15 ms 209.148.235.218
6 * * * Request timed out.
7 25 ms 17 ms 29 ms 192.178.98.195
8 32 ms 16 ms 16 ms 172.253.69.115
9 75 ms 26 ms 16 ms yyz10s17-in-f4.1e100.net [142.251.33.164]

Trace complete.
 
Everything you have posted shows no problems. Not uncommon for intermittent problems.

Pingplotter should have a required training that shows what issue really look like.

When you get loss in the middle but no loss to the end node this is testing error. Obviously you can't have 100% loss at a hop you would never see the nodes past it. There are a couple reasons you see this issue mostly the router is configured to prevent denial of service attacks so it will ignore some ping and trace packets so it doesn't impact the cpu.

It would be nice if hop 2 responded. That is the most common point of issue and represents the connection between your house and the ISP.

So pretty much it is manually do what pingplotter is doing. First run ping to the final node like 8.8.8.8 and to your router which is hop 1. You should never see delay or loss to your router. You then start to work your way up the hops. Again hop 2 would be nice. Hop 3 though is still inside your ISP network. Problem is it is hard to say if the problem is with connection to your house or something between hop 2 and hop 3.

You can keep trying higher number hops but eventually you will leave your ISP and what can you possibly do if you find the problem is in say googles network.
 
Aug 1, 2024
4
0
10
Everything you have posted shows no problems. Not uncommon for intermittent problems.

Pingplotter should have a required training that shows what issue really look like.

When you get loss in the middle but no loss to the end node this is testing error. Obviously you can't have 100% loss at a hop you would never see the nodes past it. There are a couple reasons you see this issue mostly the router is configured to prevent denial of service attacks so it will ignore some ping and trace packets so it doesn't impact the cpu.

It would be nice if hop 2 responded. That is the most common point of issue and represents the connection between your house and the ISP.

So pretty much it is manually do what pingplotter is doing. First run ping to the final node like 8.8.8.8 and to your router which is hop 1. You should never see delay or loss to your router. You then start to work your way up the hops. Again hop 2 would be nice. Hop 3 though is still inside your ISP network. Problem is it is hard to say if the problem is with connection to your house or something between hop 2 and hop 3.

You can keep trying higher number hops but eventually you will leave your ISP and what can you possibly do if you find the problem is in say googles network.
Yea hop 1 has no ping issues/PL, hop 2 has ping spikes, hop 3 is unreachable so I did another ping test to 8.8.8.8
Hop 2:
Ping statistics for 9x.xxx.xx.x:
Packets: Sent = 100, Received = 100, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 9ms, Maximum = 129ms, Average = 25ms
"ping 8.8.8.8 -n 100":
Ping statistics for 8.8.8.8:
Packets: Sent = 100, Received = 100, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 7ms, Maximum = 288ms, Average = 32ms
 
I guess its good that hop 2 will talk to you. Problem is the ISP will have issues fixing lag spikes. If you get packet loss it generally means there is something defective. Signal level issues or maybe broken equipment.

When you get latency it means data is being held in a buffer someplace. This is almost always due to some load issue. The ISP will never admit they oversold the network segment going to the houses near you.

Although it is very uncommon now days be very sure you are not using 100% of the bandwidth you buy. Most times it is upload that you exceed....or maybe microsoft is download its stupid auto update patch. Was much more a issue when people had well under 100mbps internet connections.
 
Aug 1, 2024
4
0
10
I guess its good that hop 2 will talk to you. Problem is the ISP will have issues fixing lag spikes. If you get packet loss it generally means there is something defective. Signal level issues or maybe broken equipment.

When you get latency it means data is being held in a buffer someplace. This is almost always due to some load issue. The ISP will never admit they oversold the network segment going to the houses near you.

Although it is very uncommon now days be very sure you are not using 100% of the bandwidth you buy. Most times it is upload that you exceed....or maybe microsoft is download its stupid auto update patch. Was much more a issue when people had well under 100mbps internet connections.
quick question but how would I know that I'm not using 100% of the bandwidth I buy?
 
Aug 1, 2024
6
1
15
Ex Broadband tech support person here.

I've seen this issue in the past.

Typically it has related to high upload utilisation - like bittorrent, file syncing apps like onedrive/google drive etc, where upload bandwidth is set to unlimited. Especially if these apps are monitoring folders that see regular file updates etc.

Pings/traces aren't likely to be too revealing here as they look like normal use of your connection (typically between low 50s to several hundred on the latency). I've seen it up in the thousands with high upload utilisation but that was a firmware issue where QOS wasn't working properly.

Other than that, you can have any number of potential issues causing that, it can be electrical (lookup REIN and SHINE issues) or with the cable infrastructure. An
 
quick question but how would I know that I'm not using 100% of the bandwidth I buy?
Many routers now have utilization graphs.
If you can run just your pc and no other device in the house then you can use the simple resource monitor in windows to show you all the processes that are running and how much bandwidth each is using.

It is kinda hard to overload download rates. Watching a 4k netflix movie only uses 30mbps. So if you have a 100mbps plan you could watch 3 different ones at the same time and still have enough left over for other things. Games only use about 1mbps.

You would be more likely to overload a upload rate if it is only 40 mbps. Most people don't upload anything. If you live stream than it might be a concern. What does your ISP say your upload rate should be 40mbps seems a bit low......but your average person does not use even close to that which is why they divide the total bandwidth allocating more to download.

Bolte has a very good point. Make very sure you are not using any fancy feature in your router. Anything related to QOS tends to not work well. Even other simple features like parental controls can also cause issues. Routers have very small CPU and it does not take much to spike the cpu to 100%. Normal NAT type of traffic is done in asic so the router can run 1gbit with almost no cpu load. Maybe reset the router to factory and only set the admin and wifi password. By default most those fancy features are turned off. You can of course just manually check and turn them off.