Question Ping spikes only on international servers

[bill001g]

It seems to show more of the same.....it all depends on what you normally see to google.

It is somewhat surprising you get that much latency to 8.8.8.8. Google duplicates that IP in many places and for most people it is in a large city very close....most times well under 10ms.

Unfortunately when you have so may private IP address in the path it hard to tell which ISP it is running in. My guess would be the 172.18.x.x. where you see the first large jump is your ISP.

The other nastly issue with using trace is it only shows the path to the IP the return path could be very different. ISP tend to keep the traffic on their network as long as possible. So it could your ISP international fiber to go to the site and use google international fiber to return. Not way to tell unless you can trace from the other end.

It is technically possible to get into some ISP routers and do tests if they have what is called lookglass web site.

 

[lantis3]

I logged over 7000 pings and did not have a single drop using the following command from U.S. west coast, and it's consistently around 210ms.

C:\Users\admin>ping -t www.wp.pl > c:\logs\wp.pl.log

Reply from 212.77.98.9: bytes=32 time=210ms TTL=43

 

[silver085]

It seems to show more of the same.....it all depends on what you normally see to google.

It is somewhat surprising you get that much latency to 8.8.8.8. Google duplicates that IP in many places and for most people it is in a large city very close....most times well under 10ms.

Unfortunately when you have so may private IP address in the path it hard to tell which ISP it is running in. My guess would be the 172.18.x.x. where you see the first large jump is your ISP.

The other nastly issue with using trace is it only shows the path to the IP the return path could be very different. ISP tend to keep the traffic on their network as long as possible. So it could your ISP international fiber to go to the site and use google international fiber to return. Not way to tell unless you can trace from the other end.

It is technically possible to get into some ISP routers and do tests if they have what is called lookglass web site.

Latency's to Google was high during that time because I specifically did tracert and path ping during that ~30 seconds latency spike, in hope we can finally conclude it's ISP's fault.

Here's when it's normal :

I logged over 7000 pings and did not have a single drop using the following command from U.S. west coast

C:\Users\admin>ping -t www.wp.pl > c:\logs\wp.pl.log

Reply from 212.77.98.9: bytes=32 time=210ms TTL=43

You do have other ISP, and I just checked Downdetector for my ISP, and it seems they're having problems today - https://downdetector.pl/status/vectra/

Well after all I don't think we can do much. The first problem with random high latency spikes on foreign servers seems to be gone. Now it's just my ISP having problems every now and then which affects their whole network in our country.

 

[bill001g]

The problem appears to be the router 172.17.154.1 or possibly the connection between that router and the one before it.

This almost has to be your ISP. It would be nice even if the can not fix it for you they could explain why it is happening. Not sure they will admit they are out of capacity though.

 

[lantis3]

Why did 172.17.64.10 appear twice?

Years ago my ISP seemed to have some kind of route looping between several IP addresses and it took almost a month for them to fix it.

 

[bill001g]

Why did 172.17.64.10 appear twice?

Years ago my ISP seemed to have some kind of route looping between several IP addresses and it took almost a month for them to fix it.

This is the problem when you have no access to the ISP equipment to see what they are really doing. Many times they configure the ping/trace to give you a loopback IP rather than a interface IP. Sometimes the interface does not actually have a IP and many times the interface is not even real they are running virtual router instances. Could be you pass between 2 different virtual routers on the same box. Just like servers, routers have become virtual things that make stuff massively confusing.

 

[silver085]

The first reply from my ISP was to restart my router (unplug from power supply for 15s) and 2nd reply to check cables and do a hard reset of my router. Just the usual stuff where they try to blame your equipment and convince that the issue is on your side.
I had both WinMTR for 8.8.8.8 and ping to www.wp.pl running and it was fine yesterday for the whole day.

Today :

I'll reply today with those new reports and wonder what will they come up with next. Probably offer a technician visit, yet the last time he was here in february, he told me he can only check the cables in my house and swap a router to a new one and if the issue persists, it's out of his control and he can't do more. Guess they won't admit they are out of bandwidth or something, as you said.

EDIT: Just remembered that a while ago, at the beginning of this thread, when it was only about foreign servers, I sent them one of those WinMTR reports to that MMO server that I was posting here in the beginning where packet drops and latency spikes were starting on level 3 network, and they did reply that this report doesn't indicate that there are any problems in their infrastructure. Back then when it wasn't their problem they acknowledged this. Yet now they're not commenting on current reports.