I recently discovered that a managed platform update on my AWS Elastic Beanstalk environment led to the automatic termination and recreation of my EC2 instances. Unfortunately, this process left my manually assigned Elastic IPs unassociated. I'm trying to understand how to prevent this issue in the future. Here's a bit more about my setup:
- I have an Elastic Beanstalk environment using an Auto Scaling Group with a minimum of 2 instances and a maximum of 4.
- I had Elastic IPs manually bound to specific instances, which are crucial since I use them for API Gateway resource policy whitelisting.
My main questions are: 1. How can I automatically re-associate Elastic IPs during updates? 2. Is there a way to disable these automatic updates or at least have some control over when they occur?
2 Answers
You can actually disable the managed platform updates entirely by switching to manual updates. This gives you full control over when updates occur, but keep in mind that you’ll have to manage them more actively. It's a trade-off, but it might work better for your production environment if stability is a higher priority!
It’s generally recommended to use Elastic Load Balancing (ELB) with your Elastic Beanstalk setup instead of binding Elastic IPs to individual instances. This way, your traffic will go through the load balancer, which has a static IP. But if you need to stick with Elastic IPs, you might want to set up a script that automatically re-associates them after the update. It could run as a post-deployment hook or something similar. Definitely check out the AWS documentation for the specifics on managing updates!
Thanks for the tip! I'll look into using ELB for my setup. I didn't realize that could help with the static IP issue.