Today, October 20th, we've seen another outage in AWS's North Virginia region (us-east-1), which seems to happen at least once a year. This region is the oldest and often the first to get updates, making it a testing ground for new features. While many companies choose to develop in us-east-1 for the latest features, it raises the question of reliability. Should companies keep their production environments in other, more stable regions like Ohio (us-east-2) while still developing in Virginia? Personally, I prefer the Stockholm region for its cost-effectiveness and stability, even if it lacks some of the newest features like t3a instances. Did you encounter any issues due to the outage? Our team faced some minor disruptions with Framer and Jira. What multi-region strategies do you have in place?
5 Answers
I’m currently traveling and didn't experience any major impact, just a couple of financial services hiccups. What I think I’ve learned is that us-east-1 is prone to going down more frequently than other regions. It’s definitely time for companies to consider moving away from it or adopting a multi-region strategy with redundancy.
Wait, I thought you said Virginia is the first region for feature updates? How do you know that’s true?
What if we just go for a walk until it comes back up? It's like rare events like these are perfect for some fresh air.
We developed a self-service automated failover solution years ago that doesn't rely on us-east-1. When we had to failover our critical apps today, we did it within just 15 minutes. It's crucial to coach management to act quickly since AWS updates can be slow and unreliable. If your business can't handle outages for long, I highly recommend this approach.
That's solid advice. Always have a backup plan ready—it’s not a question of if something will go wrong, but when.
Why not just implement a multi-region strategy instead?
Honestly, sometimes you just have to sit tight and wait it out. Either AWS will get things back up, or you can use the time to improve your resume.
Haha, right? They could even market these outages as a mandatory AWS Gameday!
It’s the default region where new features are rolled out. I remember a significant breaking API change that affected us-east-1. Since critical services run there, it's often more vulnerable to outages.