Question Ping spikes every 5-10 minutes ?

[Tormentula]

Almost like clockwork every 5-10 minutes, ping will spike to ~800ms for a second. It'll show up at 800 on ping plotter and sometimes cmd prompt, depends on the timing of the ping interval what number it is. Hop 1, Hop 2, and google respectively;

This has been going on for 3 months. ISP agents came over 3 times, modem was swapped out 3 different times, cables were all replaced and tested, ping plotter shows the 800ms at random points, once at hop 1 which makes no sense, sometimes hop 2, 7, 10, etc.

I tried it on a wireless laptop, had the issue... on ethernet desktop... had the issue... redirected the laptop directly to the gateway avoiding ethernet entirely, still had the issue.

Whenever those spikes occur everything freezes up, multiplayer games like league desync, and its just making everything so much more frustrating. I'll try anything but I'm pretty sure I've tried it all since the ISP doesn't seem to know what the problem is.

 

[cruisetung]

The reality is that there is nothing you can do in most cases if the glitch is outside of your home network.

If it's so critical, change ISP or use a VPN to change your route see if it helps.

 

[bill001g]

These really should not cause that much issue. 200ms is a lot but you generally see a extremely short lag it does not completely freeze or disconnect. You can completely lose 1 packet and most times the game doesn't notice.

If you would get 2 or 3 in a row that is more of a problem.

You have to be very careful about pingplotter and even the normal ping command. Pingplotter for example will insist that it getting 100% loss on say hop 4 but hop 5 and past get no loss. This is technically not possible since the data must pass through hop 4 to get to hop 5.

The issue is you have people that will use a ping command to run denial of service attacks against routers. Most routers in the internet are designed so responding to ping is low on the list for the router to process. Some also limit the total amount they will respond to in a second.

You just have to know that you get a large amount of testing error using ping.

Not sure what to suggest. Swapping out the modem is always the first thing but is seldom the issue. Many years ago there were issues with certain chipset but they were fixed and there are must faster and newer chipset being used.

Do you have just a modem or a modem/router combo unit. If it is a combo unit can you put it in bridge mode. You want to plug a pc in without any NAT function. Highly unlikely it will make any difference but you have little else to try.

It will not be a cable or signal problem. Those cause packet loss..and are much simpler to fix. Latency spikes is mostly caused by some device getting more traffic that it can process and placing the data in memory rather than discarding it. This is what you see called bufferbloat but the problem is in the ISP equipment and nothing you can do because it is your traffic combined with everyone elses traffic that is overloading something.

I am somewhat surprised the ISP even cared about this. They do not promise any kind of latency and pretty much anything related to the physical connections is packet loss not latency.

 

[Tormentula]

Its not just 200, usually what I see is this;

https://imgur.com/KnQOdBh

View: https://imgur.com/KnQOdBh


Typically I'll be playing, everything freeze up, and then check my CPU + this to see if the freeze was detected, and that's usually what it displays, sometimes it shows 200, sometimes it shows 700-800. CPU shows nothing in terms of usage spikes, and after checking 3 different devices the problem persist with all of them.

This morning I went into router settings and increased the DHCP lease on the TP link router from 120 minutes to 2880 (max), still seeing the same spikes even though everything was renewed/released but seems less frequent now? I went to my nighthawk CM2050V settings, but there's no options anywhere to customize it at all? I noticed by doing the DHCP its almost like it happens less frequently, but it still happens (perhaps its a combination of DHCP updating on the router & DHCP updating on the modem?). Still happened with the ISP's gateways before I got my own modem. I also see its not sync'd when it happens... devices individually do it at different times but this modem doesnt even have wifi so i'd assume it'd just bundle them all off the tp-link router unless im very misguided on how DHCP works.

Manually writing down when this occurs via cmd;

10:13AM - Ethernet Desktop 300ms
-- tinkered with the TP-link's DHCP lease duration around half way here.
12:03PM - Ethernet Desktop 605ms
12:19PM - MSI laptop 900ms
12:34PM - Ethernet Desktop 201ms -- 31 minutes from previous
1:05:23PM - Ethernet Desktop 677ms -- 31 minutes from previous
1:36PM - Ethernet Desktop 266ms -- 31 minutes from previous (watching for 2:07)
1:43PM - MSI laptop 726ms -- ping plotter actually didn't pick up on this while cmd did.
1:46PM - MSI laptop 130ms (probably irrelevant)
1:49PM - MSI laptop 1775ms -- new peak, didnt show on pingplotter, watching for 1:52)
1:52PM - MSI laptop 1009ms
1:52:58PM - MSI Laptop 923ms
ignoring laptop for now
2:07:21PM - Ethernet Desktop 786ms -- 31 minutes, nearly the same second mark, 2:38 surely the next.
2:23PM - Ethernet Desktop 507ms -- 16 minutes, breaks the pattern, still going to see what 2:38 does.
2:38:23PM - Ethernet Desktop 476ms -- 31 minutes from the 2:07, making 2:23 somewhat of a midway point.

There seems to be a consistency of the 31 minute runtime part, but the nighthawk modem event logs dont show anything at those time stamps while the router DHCP was moved to 2 day intervals. Just now after watching those spikes, I made the DHCP pool 1 singular ip.