2degrees latency spikes at night

Forums › 2degrees (including Slingshot, Orcon, Flip, Stuff Fibre, MyRepublic, 2talk and Vocus) › 2degrees latency spikes at night

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#318010 6-Dec-2024 10:17

Hello

I have these going on nearly every night (300/100 plan fibre):

Time is 2:05-6:00AM. Internet traffic is minimal:

I cannot find anything network internal or on the UDM Pro that would cause that. Anyone else have this phenomenon?

I'll install a separate monitor today to see if this is a reporting fluke or real.

One other issue I have that doesn't quite concern me but may be related is that on the upload I only get 90Mbit/s and not the full 100. When I reboot the ONT it will go to 100 for about a day and then drops back to 90.

Cheers Oliver

1 | 2

4582 posts

Uber Geek

ID Verified

Trusted

#3317477 6-Dec-2024 10:37

Not seeing anything like that here (UDM SE).

You just using the default latency test server in your UDM Pro?

Opinions are my own and not the views of my employer.

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3317482 6-Dec-2024 11:01

CYaBro:

You just using the default latency test server in your UDM Pro?

Yes. As I said I'll need to install something that will give me a second opinion.

Linux

11399 posts

Uber Geek

Trusted

Lifetime subscriber

#3317484 6-Dec-2024 11:10

If you are getting 90Mbps or 100Mbps per second it is not a issue / fault

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3317485 6-Dec-2024 11:44

Just remembered I have separate monitoring going (uptime-kuma ping to 1.1.1.1)

And 1 week view

SirYohnny

1 post

Wannabe Geek

#3317780 7-Dec-2024 12:12

Getting same issue,

However not only am i getting latency, i am also getting packet loss, ERR connection / taking a min to resolve Trademe and load. Happens at 5pm, happens at 5 am, a bit lost. Only thing i can think of is its ONT to ISP.

Have tried different DNS, backup router which worked fine 6 months ago, has same issue on every device, wired or wireless.

Issue started when i moved from Orcon to 2degrees. Happened day one and has happened every day (3 weeks with 2 degrees)

Tech support is the same, its your modem... its not.

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3318265 9-Dec-2024 09:43

@SirYohnny

I have seen it happen several times. But it tends to go away again. I know Chorus does regular speed tests of each connection monthly but that doesn't take that long. I also get packet loss. Haven't called 2degrees yet (it's a pain getting past support to technical).

It's gone again now.

saf

155 posts

Master Geek

ID Verified

Trusted

Vetta Group

Subscriber

#3318270 9-Dec-2024 10:00

Given the time of morning (2am), I would say this hints to provider planned works.

Looking at 2D's network status page, and searching for either "05/12" or "06/12", there's quite a lot of 2Degrees works taking place which 2am would be within.

https://www.2degrees.nz/business/network-status

My views are as unique as a unicorn riding a unicycle. They do not reflect the opinions of my employer, my cat, or the sentient coffee machine in the break room.

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

BMarquis

445 posts

Ultimate Geek

Trusted

Chorus

Lifetime subscriber

#3318278 9-Dec-2024 10:39

" I know Chorus does regular speed tests of each connection monthly but that doesn't take that long."
Chorus doesn't do that, no.

I agree with @saf this looks like 2Degrees or someone else's maintenance overnight on a couple of nights that week. Where the result was whatever you are testing to went to Aus instead of staying in NZ.

Basically the redundancy of the internet did its job.

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3318280 9-Dec-2024 10:55

@BMarquis sorry "speedtest" may be the wrong term but they do do regular line tests.

But yeah it looks like maintenance and judging from the latency that is exactly what happened. Thanks for that. It's good having the link to the change register. Didn't know they listed them out specifically. I was getting annoyed as it was starting to happen more than I was used to and every time my monitoring acts up and I start debugging my networ 😬🤣

BMarquis

445 posts

Ultimate Geek

Trusted

Chorus

Lifetime subscriber

#3318282 9-Dec-2024 11:00

olivernz:

@BMarquis sorry "speedtest" may be the wrong term but they do do regular line tests.

But yeah it looks like maintenance and judging from the latency that is exactly what happened. Thanks for that. It's good having the link to the change register. Didn't know they listed them out specifically. I was getting annoyed as it was starting to happen more than I was used to and every time my monitoring acts up and I start debugging my networ 😬🤣

Just to set the record straight - no, Chorus doesn't do regular tests of all connections. We don't do any testing like that which would impact services.

I think you are fighting a loosing battle to try and understand the cause for overnight (or even day time), short term, latency shifts.
LFCs, RSPs, IX's and content providers/CDNs are always performing maintenance or routing changes and may or may not tell end users.
e.g if cloudflare shifted 1.1.1.1 to Oz for a bit for whatever reason - you would probably never know why, might not find a record of it and you might not even notice it.

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3318295 9-Dec-2024 12:32

Ok, I'll concede. When I was working there they were running regular tests. Admittedly a while ago.

Yeah the shape shows against several different sites that get tested so unlikely to be a provider of something. And you're right, unlikely to get an answer for this but I think I have enough above to exclude my own network. Which means I can relax ;-)

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3320255 13-Dec-2024 19:03

Hmmmm.... so, called 2degrees and they checked the ONT and apparently it's the ethernet cable between the Ont and the UDM Pro. Now that'd be a 1st but there's lots going on in my 19" rack so there could be a lot of interference. I have now put ion a fat shielded cable. Let's see if that makes a diff.

CYaBro

4582 posts

Uber Geek

ID Verified

Trusted

#3320265 13-Dec-2024 19:44

There was a latency spike last night around 10pm.

Opinions are my own and not the views of my employer.

timmmay

20575 posts

Uber Geek

Trusted

Lifetime subscriber

#3320283 13-Dec-2024 20:37

olivernz:

Hmmmm.... so, called 2degrees and they checked the ONT and apparently it's the ethernet cable between the Ont and the UDM Pro. Now that'd be a 1st but there's lots going on in my 19" rack so there could be a lot of interference. I have now put ion a fat shielded cable. Let's see if that makes a diff.

I had one of those little ethernet patch cables fail once. It failed completely though, rather than increasing latency.

olivernz

497 posts

Ultimate Geek

ID Verified

Trusted

Lifetime subscriber

#3320295 13-Dec-2024 22:11

If -and that's a very big IF- the cable was the cause it won't actually be the cable but interference from powerlines etc. So the cable isn't bust just not shielded enough. But I honestly need to see it to believe it. I have never ever come across such a beast. Not even in big DCs. But I can be converted ;-)

1 | 2