Afrihost - Pure Fibre Feedback Thread

[jimini]

Yes we are seeing the increased latency from 19:00 to around 22:00 daily. Hoping it will be resolved before the weekend.

Hi @AfriNatic ,

I am also experiencing this increased latency along with 1-2 minutes of packet loss every 10 minutes or so, every single night, on EU servers (Apex Legends). Does not appear to have been resolved as hoped.

Do you have any updates on this issue?

 

[AfriNatic]

Hi @AfriNatic ,

I am also experiencing this increased latency along with 1-2 minutes of packet loss every 10 minutes or so, every single night, on EU servers (Apex Legends). Does not appear to have been resolved as hoped.

Do you have any updates on this issue?

This issue is completely resolved.

All our monitoring software shows no latency increases or loss since the issue was resolved on the 28th of April.

 

[sand_man]

 

[sand_man]

Unwatchable....

 

[mikewazar]

This is actually crazy - this is LOCAL TRAFFIC. What is going on and who is accountable?

 

[mikewazar]

Did NOC not forecast demand?
Did a supplier drop the ball?
Is there a fibre break?
Is this on Google's end? (Why is only Afrihost impacted)

 

[AfriNatic]

Unwatchable....

View attachment 1303622

Hey,

I would advise to enable IPv6 again.

 

[AfriNatic]

This is actually crazy - this is LOCAL TRAFFIC. What is going on and who is accountable?

Did NOC not forecast demand?
Did a supplier drop the ball?
Is there a fibre break?
Is this on Google's end? (Why is only Afrihost impacted)

Hi,

The issue is not on our network and not something we can control and we really want this issue to be resolved. Google has confirmed in numerous emails to us that they are aware of the capacity issues in JHB. It's unfortunate that our traffic levels causes our clients to be reroute more aggressively compared to smaller ISP's and frankly it's unfair.

We have been on Google to give us caches to wish they have finally agreed for JHB but it will take some time to arrive.

We tried reaching out to anyone that we can get capacity from and no one is able to assist. We are stuck in a sticky situation with no way out.

We aren't just sitting back and accepting our fate as we continue to work routing and other means to improve the experience.

 

[eshar]

Hi,

The issue is not on our network and not something we can control and we really want this issue to be resolved. Google has confirmed in numerous emails to us that they are aware of the capacity issues in JHB. It's unfortunate that our traffic levels causes our clients to be reroute more aggressively compared to smaller ISP's and frankly it's unfair.

We have been on Google to give us caches to wish they have finally agreed for JHB but it will take some time to arrive.

We tried reaching out to anyone that we can get capacity from and no one is able to assist. We are stuck in a sticky situation with no way out.

We aren't just sitting back and accepting our fate as we continue to work routing and other means to improve the experience.

wont NAP be able to help while waiting for google ?

 

[AfriNatic]

wont NAP be able to help while waiting for google ?

That is where the issue is. We are not sure if it's peering capacity or further on Google's network that is the issue but one of our suggestions was to fire up some JINX peering but Google advised that JINX does not have extra capacity over NAP.

 

[Sicilian-Najdorf]

Hi Guys
I came across these little UPS devices used to keep you connected during load shedding by powering up the router and ONT.
How well do these devices work?

Does Afrihost have the infrastructure to remain online during load shedding?
It's pretty pointless to get one of these devices if the underlying infrastructure is not resilient.

Area: Kraaifontein - Northern Suburbs - WC

EDIT: Infrastructure is done by Octotel

 

[KillerOnDrugX]

Hi Guys
I came across these little UPS devices used to keep you connected during load shedding by powering up the router and ONT.
How well do these devices work?

Does Afrihost have the infrastructure to remain online during load shedding?
It's pretty pointless to get one of these devices if the underlying infrastructure is not resilient.

Area: Kraaifontein - Northern Suburbs - WC

Works well. I have the UltraLan Micro UPS just for my ONT and router and it can go for like 16 hours of backup if my electricity is out. Depends on the FNO I believe whether there is backup power or not. Openserve seems to always have backup but on that note, Afrinatic will have to advise. Who is your FNO?

 

[Sicilian-Najdorf]

Works well. I have the UltraLan Micro UPS just for my ONT and router and it can go for like 16 hours of backup if my electricity is out. Depends on the FNO I believe whether there is backup power or not. Openserve seems to always have backup but on that note, Afrinatic will have to advise. Who is your FNO?

Octotel did the infrastructure

 

[Hook]

I'm also using the UltraLan Micro UPS. Easily lasts through load shedding and charges quickly. I'm on Metrofibre FNO and it has so far always stayed up during loadshedding.

 

[Vorastra]

That is where the issue is. We are not sure if it's peering capacity or further on Google's network that is the issue but one of our suggestions was to fire up some JINX peering but Google advised that JINX does not have extra capacity over NAP.

Whatever the issue is, it seems to be affecting WebAfrica too the last few days in the evening. Mostly it's the livestreams that suffer the most. I've been keeping an eye here since you're so active and the issue is also affecting Afrihost.
I saw some people saying that it was a Durban issue specifically.

I notice you ask people to check googlevid director. Mines at echoza-cpt1. Where are you guys rerouting to?

 

[ZyborgRSA]

Anyone else in Durban getting random disconnects/Massive spikes?

 

[bradza]

Anyone else in Durban getting random disconnects/Massive spikes?

At one time I even had international sites state my IP was banned from accessing that site (and couldn't access others either, legit company sites) when it shouldn't ever be saying that. It's as if our IP's were bad or something. It unlocked after a short while while in contact with my support in Los Angeles and they couldn't think of anything as it wasn't their end, so seems something strange locally.

 

[SilverCode]

Anyone have problems with Amazon Prime Video in the evenings? During the day it works 100% fine. Around 7pm it fails to load and says there are "Internet Connectivity Problems". This happens on both my streaming devices, and it is only ever at night.

*Edit: Forgot to mention, other streaming services continue to work mostly fine (Netflix, YouTube, Apple TV+, BritBox, Showmax).

 

[bradza]

Youtube can't even stream this side currently, Durban. Even 180p buffering. Just a heads up.

 

[JustJack]

Once again despite every setting being in place, YT just an absolute sh!tshow