Hi all,
When deploying an application in a private cloud (Kubernates / Openshift), we have to configure the Core resources of the application (see https://docs.mendix.com/developerportal/deploy/private-cloud-deploy/#create-environment for more information).
Mendix Portal has limitation for those resources, in particular, it is not possible to set CPU request lower than 0.1 CPU.
This CPU request will reserve CPU resources for the application. It means that no other application in the namespace can use this resource if needed. CPU request is dedicated to the application.
On the other hand, CPU limit is the maximum amount of CPU that an application can use, if available in the namespace. So if the app need computing resources it will pick it in the shared CPU configured in the namespace.
At my company, when we analyze CPU consumption, we see that most of the time, the real consumption is between 0.01 and 0.05 CPU which is 10 to 2 times lower than the requested value (0.1) that we have to set due to Portal limitation.
When we pay our Kubernates / Openshift licences based on the total amount of CPU we have in the cluster, it is important to optimize the resources.
So, my idea is: allow Mendix for private cloud customers the set CPU request at a lower value than 0.1 CPU, at least till 0.01 CPU.
0.1 CPU should only be a recommended value.
It is then up to us to set the value we want and if we are below the recommended value we will take responsibility if we run into problems.
Thanks.