if I have a deployment with 3 replicas running on Nginx and I set a memory request for 500Mi (memory) and 200m (CPU)
So does that mean, I should have 1500Mi (memory) and 600m (CPU) vacant in my node for the deployment to run it can run with fewer resources as well?
Good question @pikachunetes . I think you are right. however before you do your Request, you have to check the Limit configured and make sure your Request won’t exceed the Limit values. Over all Server CPU & memory matters
I think if it will a limit instead of a request then deployment would be able to run.
Shwetha Shenoy V:
I believe all 3 pods of the deployment would start with these request on a node with 1500Mi mem and 600m CPU. If Limits arent set(to be same as Request value) and if the pods try to “grow”, they wont have enough resources and you may see pod/container restarts due to OOMKiled or CPU throttling.
Its a good practice to set both Limit & Resource to avoid any resource scarcity