Web Squad ISP

[Zodiak]

Thanks, let me see if we can try route around JINX for an A/B test. I've asked the team to drain JINX; please test again in the next 5-10 mins. Will keep digging on this side. Let's see if JINX change / surfshark helps and then troubleshoot from there. Not that there's anything wrong with JINX (we currently see about 15-25% of our traffic carrying over this point without issues and we are big fans of JINX's JHB fabric) but this would assist with narrowing down potential bottlenecks upstream. Draining this one point just moves traffic to other points and helps speed up the diagnostic process.

Update: JINX is drained

Running smooth like butter for now, 0% packet loss or even packet bursts!

 

[websquadza]

Running smooth like butter for now, 0% packet loss or even packet bursts!

Great news, thanks for the update. Given that two changes were made, let me ask the team to fold JINX back in (we prefer having all services running). Please update if you notice any changes.

 

[websquadza]

Ok so something new and interesting is happening, Speedtest to any server in London is between 300 - 370ms ping and terrible download speed, the rest of EU is fine though and US is fine.....

Ok so something new and interesting is happening, Speedtest to any server in London is between 300 - 370ms ping and terrible download speed, the rest of EU is fine though and US is fine.....

Noticed this and our graphs corroborate; I think one of our transit vendors advertised some of our prefixes to US-EAST resulting in traffic routing that way. This has been resolved, though we have still reached out for a confirmation.

 

[websquadza]

I'm in CPT, not seeing anything to that effect:

Tested to Vodafone UK

I'm seeing pings in Apex to EU servers are generally higher these days.
Mine regularly sits around 210 - 250 ms, while friends are around 170 to 200 ms...

We are working on BGP communities with our vendors to better steer inbound traffic, to prevent this higher latency. Some EU return paths favour a specific JHB transit provider. We obviously don't want to drain advertisements for redundancy reasons.

 

[Zodiak]

Great news, thanks for the update. Given that two changes were made, let me ask the team to fold JINX back in (we prefer having all services running). Please update if you notice any changes.

Will definitly test tonight and will advise if the packet loss is back.

 

[Zodiak]

Great news, thanks for the update. Given that two changes were made, let me ask the team to fold JINX back in (we prefer having all services running). Please update if you notice any changes.

So I did a couple of test matches and unfortunately I am seeing packet loss/packet bursts again but definitly not as severe as I saw over the weekend but not completely gone

 

[Zodiak]

Ok did a few more and so far still ok, no major packet loss, will keep testing through the week

 

[websquadza]

So I did a couple of test matches and unfortunately I am seeing packet loss/packet bursts again but definitly not as severe as I saw over the weekend but not completely gone

Local servers?

Ok did a few more and so far still ok, no major packet loss, will keep testing through the week

Let me know, will get the guys to look at JINX interface in the meanwhile

 

[Zodiak]

Local servers?

Let me know, will get the guys to look at JINX interface in the meanwhile

So far looking better over the last few days, will keep monitoring

 

[Zodiak]

Still rock solid, not sure what changed but looking good

 

[GoneWithTheWindows]

@websquadza - getting some REALLY annoying loss - ping is up and down every few seconds....

 

[websquadza]

View attachment 1526331
@websquadza - getting some REALLY annoying loss - ping is up and down every few seconds....

Looks like last mile there, packet loss is between your router and our BNG. If this doesn't clear up within an hour or so, please log a fault so we can get on the FNO's case.

 

[PsyWulf]

View attachment 1526331
@websquadza - getting some REALLY annoying loss - ping is up and down every few seconds....

1% from your router all the way to goog. That's your line

 

[jvrs]

Any known Octotel outages in Muizenberg ?

 

[jvrs]

Any known Octotel outages in Muizenberg ?

@websquadza Ticket ref# 246826

 

[DarkSt0rm]

Looks like vuma has shat the bed lol, the CISP thread is showing the same issues lol







Things have been fine since the SFP replacement

 

[websquadza]

Looks like vuma has shat the bed lol, the CISP thread is showing the same issues lol

View attachment 1527363


View attachment 1527365
View attachment 1527367

Things have been fine since the SFP replacement

Morning, you still seeing these issues?

Looking at the logs I see the same- trying to isolate a little further as there was some weirdness on some unrelated parts of the network too. We’re just digging a little deeper to make sure there isn’t something common.

Thanks for the update on the sfp!

 

[feo]

I've sent a couple of PMs but no reply as of yet.

I'm interested in a test account, please.
I'm with Vox and having latency issues connecting to my work VM over Citrix.

Vox support is terrible so I need to test using another ISP account to see if it resolves my issue.

 

[websquadza]

I've sent a couple of PMs but no reply as of yet.

I'm interested in a test account, please.
I'm with Vox and having latency issues connecting to my work VM over Citrix.

Vox support is terrible so I need to test using another ISP account to see if it resolves my issue.

Morning
thanks for your follow up - I have responded on your PM

 

[feo]

Morning
thanks for your follow up - I have responded on your PM

Thank you, and back at you.