Gregory's ServerStatement regarding the ongoing SourceHut outage
https://outage.sr.ht/
 7 comments
@jakintosh yeah, it's just pain to have to change all the links to point to the new remotes. I think we rely a bit too much on sr.ht for code distribution, we're looking to make our shipping code a bit more robust and decentralized now, it's a good wake up call.  @neauoire thanks for posting. I'm AFK and largely unaffected. But as a customer it's useful to follow. @neauoire I also experienced a hosting provider null routing a service (without telling me) in similar circumstances. My response was to move to a new provider... A truely competent provider should be able to cap bandwidth in such a way that other services aren't affected by the DDOS I should hope...  @timthelion @neauoire I think they are reaching for simple solutions unless you pay and I'm guessing that would be very expensive. What I'm confused about is why so few people are building their systems distributed, specially if they are based off git (probably mercurial is to blame)?! Then you can have (code size allowing) hundreds of small servers everywhere. The only mitigation for DDoS is distributed hosting? Decentralized would be the dream. @timthelion @neauoire How? The only mitigation for a layer 3 DDoS commonly on the market is "I have so many resources you can't saturate them," which is why it's so expensive. |
@neauoire wow, it’s pretty demoralizing to see how someone with enough resources can seemingly nuke free software resources into dust. guess it’s a reality check though, and at least git is trivial to point to a different remote.