Here is something little tricky that I learned which is not really clear in the AWS documentation.
Problem
When you have a Kubernetes cluster in AWS (EKS), you will need to have a lot of space for each node's disk.
If not you will have some PLEG is not healthy errors.
Solution
Following advices of the AWS support, (in my case) I need to set this value to 150Go to let some space for all your pods/images/deployments...
The weird thing is that I haven't seen any documentation which explain how much we need to define for each kind of EC2 instance.
If you have this link, please send it to me.
Monitoring
Also you can monitor some values in Cloudwatch (if you have a cloudwatch agent on your cluster) to check if everything is alright, and set up some alerts to be aware if something wrong is happening.
Link : AWS Documentation
I hope it will help you!
Oldest comments (0)