Details
-
Feature Request
-
Resolution: Done
-
Undefined
-
None
-
None
-
None
-
False
-
None
-
False
-
Not Selected
-
0
-
0%
Description
1. Proposed title of this feature request
2. What is the nature and description of the request?
The min CPU request made in the deployment config is 100mi and the max limit is can upto 8 CPUs.
If there is no resource request in the deployment config, it will set with default request and default limit.
I would like to know how do I restrict "CPU requests only to a certain value (eg 500mi) in the deployment config with the above configuration.
3. Why does the customer need this? (List the business requirements here)
1. There are two different teams involved in this. One is the application team and one is the platform team. 2. There is a Quota declared by the platform team for a specific namespace. 3. Use Case: The CPU limit is set to 4 CPUs. While the application team deploys the application the amount of CPU request can be any value till 4 CPUs as the limit is 4 CPUs. 4. Let's assume that the application team requests 3 CPUs in their deployment config But when we check the actual CPU consumption for the same application it hardly uses 1 CPU so the gap between the request and actual consumption is causing inefficient resource usage. The extra 2 CPUs are not getting used also it won't be available for other pods as it is getting reserved by the deployment config. 5. We are looking for a solution that will restrict the number of CPU requests to a specific value for a deployment config let's say 1 CPU and if any deployment config requests more than 1 CPU the deployment config should throw an error and they will contact the platform team to increase the request for them. 6. We cannot set hard limits for a namespace for 1 CPU because the application team is paying for 4 CPUs and we are also good to provide 4 CPUs unless it is being utilized fully.
4. List any affected packages or components.
Node resource consumption.