Github: Major Service Outage • Hacker News Board

Re: Github: Major Service Outage

#31

Post by zzleeper » Thu, Oct 18, 2012, 8:59 PM UTC

Jgrubb wrote: post #3

I think it's probably my fault. Github seems to go down every time I start poking around in someone else's projects.

By any chance, did you 'poke' with a cluster of LOICs?

Re: Github: Major Service Outage

#32

Post by gojomo » Thu, Oct 18, 2012, 9:00 PM UTC

adgar2 wrote: post #24

bgentry wrote: post #21

Earlier quoted context omitted.

They were already running their status site on Heroku prior to the last major service outage.

This is correct. They needed ~90 dynos to keep up with about 500 qps on their status page in the last outage. Source: http://news.ycombinator.net/item?id=4524489

Should probably throw a Fastly or CloudFlare cache in front of a single dyno.

Re: Github: Major Service Outage

#33

Post by FuzzyDunlop » Thu, Oct 18, 2012, 9:02 PM UTC

Nothing against Github but this probably highlights the real benefit of DVCS: setting up multiple remotes for your repo. Manage it probably and when one service goes down, fall back to Bitbucket or another service.

It would limit the potential damage these attacks could cause, given the reliance dev teams have on pushing code to a central repo. Taking down a site like Github has a fairly clear effect on the productivity of a lot of their users.

Re: Github: Major Service Outage

#34

Post by zheng » Thu, Oct 18, 2012, 9:05 PM UTC

First off, as others have said, disabling port 80 is a great way to handle this, I don't really care that much if I can see/use github the website for a few hours, but I'd be much more upset if I couldn't pull my code.

Secondly, I kind of like when big sites go down when I'm not in desperate need because it means a really nice aftermath write up is on the way. Can't wait to hear more about this one.

Re: Github: Major Service Outage

#35

Post by brokentone » Thu, Oct 18, 2012, 9:08 PM UTC

GitHub, YouTube, I'm having trouble on Wordpress.org, anything else down for people? Anyone noticing any correlations?

Re: Github: Major Service Outage

#36

Post by alook » Thu, Oct 18, 2012, 9:09 PM UTC

creativityhurts wrote: post #5

They said they're under a DDoS attack https://twitter.com/github/statuses/259029493669310464

Who DDOS'es twitter? Mercurial committers? the Bitbucket people?

Re: Github: Major Service Outage

#37

Post by longdivision » Thu, Oct 18, 2012, 9:10 PM UTC

https://twitter.com/AnonymousOwn3r is taking credit for attacks against YouTube and GitHub amongst others.

Re: Github: Major Service Outage

#38

Post by Stefan_H » Thu, Oct 18, 2012, 9:17 PM UTC

alook wrote: post #36

creativityhurts wrote: post #5

They said they're under a DDoS attack https://twitter.com/github/statuses/259029493669310464

Who DDOS'es twitter? Mercurial committers? the Bitbucket people?

did you mean "who DDOS'es Github"?

Re: Github: Major Service Outage

#39

Post by jamesjguthrie » Thu, Oct 18, 2012, 9:18 PM UTC

Pfft, always down these days. BitBucket for me.

Re: Github: Major Service Outage

#40

Post by bluesaunders » Thu, Oct 18, 2012, 9:20 PM UTC

FuzzyDunlop wrote: post #33

Nothing against Github but this probably highlights the real benefit of DVCS: setting up multiple remotes for your repo. Manage it probably and when one service goes down, fall back to Bitbucket or another service. It would limit the potential damage these attacks could cause, given the reliance dev teams have on pushing code to a central repo. Taking down a site like Github has a fairly clear effect on the productiv…

It'd be glorious if one service (say GitHub) had multiple remote links that dynamically switched for you when service faced an outage.