Conversation
Notices
-
Embed this notice
@coolboymew @gabriel Every time that US-EAST-1 has gone down, Almost all of the other AWS regions had disruptions because so much stuff even internally relies on US-EAST-1. Supposedly they have addressed this since last time but we won't know until it happens again.
I can't remember what it was but there was a thing a few years ago that knocked out some big DNS servers and nearly everything stopped working. even if you didn't need DNS to get to your service, that service relied on DNS somehow and went down.
I'm told you could make most of the Internet unusable by knocking out one or two backbones because it's not super decentralized anymore but I don't know the details, just that most of the traffic goes through the USA.
-
Embed this notice
@coolboymew @gabriel btw if you are willing to spend the money and you design your stuff right, you can automatically replicate everything in your AWS project into another region and just switch it over in a couple minutes if your primary region goes down.
-
Embed this notice
@Moon @gabriel @coolboymew godaddy went down at some point and I didn't get an alert because either pagerduty or twilio had their nameservers on godaddy 😀
-
Embed this notice
@m0n5t3r @gabriel @coolboymew AWS availability dashboard was hosted on AWS and that failed catastrophically in the last outage so they now have moved their dashboard off AWS.