4 ms·

Network Performance Issues in Europe

40 points by owenwil 10y ago

flashm 10y ago
What's interesting about this is that Cloudfare seem to duck a big chunk of the flak - everyone immediately suspects their own server, as cloudfare gives a pretty generic error page. Lots of people chasing Heroku on twitter currently, for example when it's not their fault at all.
- dbbk 10y ago
  Really? The Hacker News error page said CloudFlare could not be reached.
shdon 10y ago
"down across Europe" seems a bit of an overstatement. Not having any problems with any CF-dependent sites from here in the Netherlands. It also seems from the status page that only Moscow is currently being rerouted. It could be teething troubles, as that datacenter is the latest to be added to their network.
- piva00 10y ago
  Here in Sweden it was quite bad for almost an hour (around 14:00 CEST), no way to access a bunch of documentation, Slack off, connection to AWS down, everything seems to be fine again since around 15:10 CEST.
- danielhunt 10y ago
  A number of high-profile, consumer-centric sites/services were completely offline for around an hour throughout what appears to be a large portion of Europe. Seems like a fair statement to me (Dublin, Ireland - a heap of sites and service were completely offline, including hackernews, whatsapp, and reddit)
modoc 10y ago
Not be a CF apologist, but CloudFlare status shows their issues pretty honestly and openly. In contrast Akamai has frequent issues (usually isolated to a small area) and virtually never admits it unless you can PROVE it's them.
xchaotic 10y ago
It's really good that they were honest about it. An hour or so of downtime basically throws any five nines SLA out of the window for the entire year.
aroch 10y ago
Seems to be a backbone / fiber issue. Leaseweb also experienced quite a large drop in connectivity and transatlantic latency went through the roof[1] [1] http://leasewebnoc.com/en/networkstatus/latency-and-packet-loss-eu-us http://leasewebnoc.com/en/networkstatus/latency-and-packet-l...
admiun 10y ago
I always wonder if there's easy way to know where current global internet interruptions are, anyone know? In this case it apparently was Telia [1] but how do you figure that out? Top Google results for 'global internet status' aren't really usable. [1] http://www.techweekeurope.co.uk/cloud/cloud-management/cloudflare-suffers-european-outage-194020 http://www.techweekeurope.co.uk/cloud/cloud-management/cloud...
- XioNoX 10y ago
  This helps a bit: http://internetpulse.net/ http://internetpulse.net/
- niftich 10y ago
  Realistically, this isn't solvable due to information asymmetry. Any one company only knows if their services are up, but doesn't really know what's going on upstream or downstream of them. That's why tracing an incident like this takes time, and most people blame the wrong party at first until the true cause is found. But Twitter has become the de facto platform where you announce stuff to people, or shout into the ether saying 'is Cloudflare down?' Signal-to-noise ratio isn't even that low -- most baseline chatter is noise, but there should be a MASSIVE, near-simultaneous spike when a high-profile site goes down.
DanielDent 10y ago
As Cloudflare continues to grow their POP footprint, the frequency of issues has also gone up. When I say issues, I mean "something, somewhere in the world is broken". More POPs and more peers means more outages. This is probably especially true for their POPs in less developed regions of the world. Cloudflare also has a team monitoring for issues and it's usually pretty easy for them to work around the issues that come up. Their design makes it so that even the loss of an entire POP isn't a big deal. As their service continues to get better and more reliable, their status page will actually look like it's getting worse and worse. But that's because they are being open and transparent about day-to-day operational details.
nallerooth 10y ago
For those of you who missed it: http://www.theregister.co.uk/2016/06/20/telia_engineer_blamed_massive_net_outage/ http://www.theregister.co.uk/2016/06/20/telia_engineer_blame...

40 points by owenwil 10y ago

What's interesting about this is that Cloudfare seem to duck a big chunk of the flak - everyone immediately suspects their own server, as cloudfare gives a pretty generic error page. Lots of people chasing Heroku on twitter currently, for example when it's not their fault at all.

dbbk 10y ago
Really? The Hacker News error page said CloudFlare could not be reached.

shdon 10y ago

"down across Europe" seems a bit of an overstatement. Not having any problems with any CF-dependent sites from here in the Netherlands. It also seems from the status page that only Moscow is currently being rerouted. It could be teething troubles, as that datacenter is the latest to be added to their network.

piva00 10y ago
Here in Sweden it was quite bad for almost an hour (around 14:00 CEST), no way to access a bunch of documentation, Slack off, connection to AWS down, everything seems to be fine again since around 15:10 CEST.
danielhunt 10y ago
A number of high-profile, consumer-centric sites/services were completely offline for around an hour throughout what appears to be a large portion of Europe. Seems like a fair statement to me (Dublin, Ireland - a heap of sites and service were completely offline, including hackernews, whatsapp, and reddit)

modoc 10y ago

Not be a CF apologist, but CloudFlare status shows their issues pretty honestly and openly. In contrast Akamai has frequent issues (usually isolated to a small area) and virtually never admits it unless you can PROVE it's them.

xchaotic 10y ago

It's really good that they were honest about it. An hour or so of downtime basically throws any five nines SLA out of the window for the entire year.

aroch 10y ago

Seems to be a backbone / fiber issue. Leaseweb also experienced quite a large drop in connectivity and transatlantic latency went through the roof[1] [1] http://leasewebnoc.com/en/networkstatus/latency-and-packet-loss-eu-us http://leasewebnoc.com/en/networkstatus/latency-and-packet-l...

admiun 10y ago

I always wonder if there's easy way to know where current global internet interruptions are, anyone know? In this case it apparently was Telia [1] but how do you figure that out? Top Google results for 'global internet status' aren't really usable. [1] http://www.techweekeurope.co.uk/cloud/cloud-management/cloudflare-suffers-european-outage-194020 http://www.techweekeurope.co.uk/cloud/cloud-management/cloud...

XioNoX 10y ago
This helps a bit: http://internetpulse.net/ http://internetpulse.net/
niftich 10y ago
Realistically, this isn't solvable due to information asymmetry. Any one company only knows if their services are up, but doesn't really know what's going on upstream or downstream of them. That's why tracing an incident like this takes time, and most people blame the wrong party at first until the true cause is found. But Twitter has become the de facto platform where you announce stuff to people, or shout into the ether saying 'is Cloudflare down?' Signal-to-noise ratio isn't even that low -- most baseline chatter is noise, but there should be a MASSIVE, near-simultaneous spike when a high-profile site goes down.

DanielDent 10y ago

As Cloudflare continues to grow their POP footprint, the frequency of issues has also gone up. When I say issues, I mean "something, somewhere in the world is broken". More POPs and more peers means more outages. This is probably especially true for their POPs in less developed regions of the world. Cloudflare also has a team monitoring for issues and it's usually pretty easy for them to work around the issues that come up. Their design makes it so that even the loss of an entire POP isn't a big deal. As their service continues to get better and more reliable, their status page will actually look like it's getting worse and worse. But that's because they are being open and transparent about day-to-day operational details.

nallerooth 10y ago

For those of you who missed it: http://www.theregister.co.uk/2016/06/20/telia_engineer_blamed_massive_net_outage/ http://www.theregister.co.uk/2016/06/20/telia_engineer_blame...