I'm having trouble deploying anything in our ECS Fargate cluster in the US-EAST-1 region. Although there have been some improvements, we're currently facing capacity issues and can't launch tasks. AWS noted that while most operations are recovering from a DNS problem, requests to launch new EC2 instances are still encountering errors. As of now, SQS isn't properly communicating with Lambdas, and tasks are failing to start due to network interface provisioning issues. Has anyone else experienced this, and do you have any suggestions?
4 Answers
It’s interesting how all this peaked when a lot of users came online. It seems like AWS wasn’t designed for that surge all at once. In the past, when we faced similar bottlenecks, staggering requests worked, but with this scale, we might need more geographical distribution to handle the load effectively.
Absolutely. I tried launching EC2 instances without specifying an AZ, thinking it might help with capacity issues. Found that it could potentially avoid some error rates. Just keep in mind that there’s still a backlog they're dealing with and it might take some time for everything to stabilize.
Yeah, I've noticed the same issue. It seems like the earlier reports of normalcy from AWS don’t hold up for everyone. Sometimes it takes a while for everything to catch up, especially under high demand. Just gotta hang tight and keep checking for updates!
It's been frustrating for sure. I noticed the same problem with Lambdas not receiving messages from SQS. They really need to prioritize those polling delays; it’s impacting several AWS features. I'm hoping they’ll resolve these lingering issues soon.
Good tip! I hadn't considered that but it makes sense. Flexibility with AZs could definitely help smooth out the process.