Help Center/ Cloud Container Instance (CCI)/ User Guide/ Using CCI with CCE/ Rounding Up Pod Resources
Updated on 2025-08-12 GMT+08:00
View PDF
Share

Rounding Up Pod Resources

The CCE Cloud Bursting Engine for CCI add-on can select appropriate resource specifications based on the vCPUs and memory of the workloads to improve resource utilization. This section describes how appropriate resource specifications are selected.

Constraints

  • At least one container in the pod that is scheduled to CCI must have vCPU and memory limits or requests specified or the resource.cci.io/pod-size-specs annotation.
  • Pods that are scheduled to CCI 2.0 must comply with the flavors of CCI 2.0. If a pod does not comply with the flavors of CCI 2.0, an error similar to the following will be reported in the pod event on CCE, and the pod will not be created on CCI 2.0. For details about the flavors, see Upgrading Pod Flavors.

How Resource Specifications Are Calculated

Some necessary system components required by the pods occupy system resources. There is a difference between the memory in the pod flavor and the allocatable pod memory. CCI calculates the pod memory that can be allocated as follows:
  • Memory in the pod flavor ≤ 2 GiB

    Allocatable pod memory = Memory in the pod flavor

  • Memory in the pod flavor > 2 GiB

    Allocatable pod memory = Memory in the pod flavor – Memory reserved for CCI – Memory reserved for the OS

    Memory in the pod flavor is the value of ${memoryCeil} displayed in the pod annotation in the format of resource.cci.io/size=${cpuCeil}_${memoryCeil}.

For details, see Reserved System Overhead.

Reserved System Overhead

For pods that are scheduled to CCI 2.0, you can also use the resource.cci.io/memory-reservation and resource.cci.io/memory-burst-size annotations in pod/podTemplate to enable system overhead reservation. For details about the rules, see Reserved System Overhead.

Parent topic: Using CCI with CCE