So I've been working with EC2 instances in two separate contexts. In the first context I simply am building an AMI for some spot instances. In the other context I have an EKS cluster with node groups.
One day everything just stopped working. I looked into it and it has to do with outbound requests from the instances. I could use ping google.com but not curl google.com for example. I created many different instances in different availability zones, different sizes, and different distros but all with the exact same network settings.
The takeaway is that Amazon Linux 2 poses no issues for outbound requests but Amazon Linux 2023 does. Are there known causes for this? Notable that this coincided relatively with the release of (more so with my update to) AMI release version 1.27.4-20230825.
Finally, how can I fix this in my EKS cluster. I don't really have a way to change distros so ideally there would be an easy workaround.