Just tested LLVM with an http download. Getting about 60-70 mbps. Can you check again?
Looks like a major outage or issue on a JHB-CPT route. We’ve failed over to another route. But noticed that JHB peers still showing loss, a few are showing improvement already. There are 3 decent JHB- CPT routes- so lots of providers share the same affected route. We’ve dropped AWS in JHB (their route was showing losses there). Running off CPT and on our transit now, so PUBG should be back to normal.Connection not doing great again tonight, buffering is back and had major packet loss in PUBG just now up to 50%
China Telecoms have partnered up with Liquid Telecoms for a transit share agreement of sorts. This meant we were receiving a fair whack of inbound traffic via Liquid (and we’re not a fan of their higher latencies). We’ve dropped China telecom until they sort this out. The wheels at CT move slowly, so we should be back up later during the coming week.Is it just me or is the latency to Asia/Singapore back to 300+ ? Last week or the week before it was sub 200ms. Anyone else experiencing the same or was there a change?
Looks like the liquid / china tel swap is causing issues throughout. And they did it without a way for ISPs to opt out or filter out. Most transit providers use communities to let you decide- these guys didn’t. This looks like the forward route is the liquid route, and return is the China telecoms (hence the lower latency). The one above shows liquid following the same route as HE and picking up flagtel in Dubai (which has always been a bad exchange for flagtel- IP ranges look similar). This is why we’ve turned them down for the time being- routing over the combination is worse than each individually- neither has tried to optimise their routes yet.