Non-user dependent Resource Usage Limiting for workloads

The Resource Usage Limiting feature enables the utilization of CPU and memory independent of the user, in a way that no resource is unnecessarily blocked.

If the Resource Usage Limiting feature is enabled, Administrators can define the standard limits of CPU and memory resources applicable to any started workload, that is, to any started job or session. When a user starts the workload, the standard resource limit set by the administrator will be used for the workload pod’s resource requests. However, the workload pod resource limits will be set based on the Resource Profile selected by the user when starting a workload.

Example for enabled Resource Usage Limiting feature🔗

If the Administrator configures the standard resource limit in 100m CPU and 0.25 GiB memory, the user will be able to create the workload using this configured standard resource, that is 100m CPU and 0.25 GiB memory. However, the resource limits used for workload pods will be the resource profile selected, that is 8 CPU and 16 GiB memory.

Example for disabled Resource Usage Limiting feature🔗

If the Resource Usage Limiting feature is not enabled, the user can create a workload with the capacity configured in the Resource profile, that is, for example, 8 CPU and 16 GiB memory. In this case, even though the workload uses only 200 m CPU and 0.5 GiB memory, it will block the whole capacity allocated in the resource profile. The unused resources in that case are blocked by the workload.