为何突然无法在AWS Elastic Beanstalk中重新部署应用程序的可能原因？

Question

I have had a PHP/RDS application running in AWS Elastic Beanstalk for about a year. Last code update was more than three months ago. A couple of days ago, the application became unresponsive, and was restarted automatically. This has worked in the past, but this time the application failed to re-launch every time. ELB is stuck in an infinite re-deployment loop. Every time a new EC2 instance is launched, it is immediately considered unhealthy and is terminated. I have tried to look at the log files, but there is nothing interesting there (as far as I can see). It's the log of installing a new EC2 instance, and then immediately switching it off.

I hadn't made any changes to my application code or the EBS settings for over three months when this happened out of the blue.

I have now tried creating a new EBS environment from scratch for the same application, but the problem remains. Has anyone had any similar experience? Have there been changes in the Amazon environment that could cause this behaviour? Any ideas or pointers would be greatly appreciated.

Answer 1

AWS had updated their EBS system. There was a new design and new options in the management console. When creating a new EBS environment, I was given the choice to use the new "beta" version, which I finally did. And from that point onwards everything has worked flawlessly.

When this happened I also wrote to AWS support. They replied eventually, not with a solution or an explanation, but with a deduction of a couple of months fee, which I interpret as a way of saying "sorry that something went wrong with your setup when we updated our servers". Just guessing on that point, though.

All in all, the new setup has worked much better than the previous one, which had downtimes every now and then.