Assigning resources to kubernetes pod

Kubernetes
#1

I am trying to assign the following resources to a pod, using " resource " attribute.

Q) CPU limits in pod.yaml is 0.2m, but when i describe the pod or node , in the output the CPU limits it is displaying 1m and not 2m.
Kindly explain.

# pod.yaml

apiVersion: v1
kind: Pod
metadata:
  creationTimestamp: null
  labels:
    run: nginx
  name: nginx
spec:
  containers:
  - image: nginx
    name: nginx
    ports:
    - containerPort: 80
    resources:
      requests:
        cpu: 0.1m
        memory: 200M
      limits:
        memory: 400M
        **cpu: 0.2m**
  dnsPolicy: ClusterFirst
  restartPolicy: Always
status: {}

# output of the command " K describe po nginx "
Limits:
      **cpu:     1m**
      memory:  400M
    Requests:
      cpu:        1m
      memory:     200M


# output of the command " k describe node <nodename> "

  Namespace                   Name                                        CPU Requests  CPU Limits  Memory Requests  Memory Limits  Age
  ---------                   ----                                        ------------  ----------  ---------------  -------------  ---
  default                     nginx                                       1m (0%)       **1m (0%)**     200M (2%)        400M (4%)      4m36s
  kube-system                 calico-kube-controllers-658d97c59c-qv9j2    0 (0%)        0 (0%)      0 (0%)           0 (0%)         18d
  kube-system                 calico-node-v6q8k                           250m (12%)    0 (0%)      0 (0%)           0 (0%)         18d
  kube-system                 kube-proxy-nn96k                            0 (0%)        0 (0%)      0 (0%)           0 (0%)         18d
Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  Resource           Requests    Limits
  --------           --------    ------
  cpu                251m (12%)  1m (0%)
  memory             200M (2%)   400M (4%)
  ephemeral-storage  0 (0%)      0 (0%)
  hugepages-1Gi      0 (0%)      0 (0%)
  hugepages-2Mi      0 (0%)      0 (0%)

Thank you
Deepak

#2

This is addressed in the docs:

Note:

Kubernetes doesn’t allow you to specify CPU resources with a precision finer than 1m or 0.001 CPU. To avoid accidentally using an invalid CPU quantity, it’s useful to specify CPU units using the milliCPU form instead of the decimal form when using less than 1 CPU unit.

For example, you have a Pod that uses 5m or 0.005 CPU and would like to decrease its CPU resources. By using the decimal form, it’s harder to spot that 0.0005 CPU is an invalid value, while by using the milliCPU form, it’s easier to spot that 0.5m is an invalid value.

1 Like
#3

Kubernetes understands the unit u for micro, thus 200u = 0.2m, however I believe the minimum amount of CPU you can assign is 1m and that’s defined by the Linux kernel, so anything less than that will be rounded up.

1 Like
#4

Also I found the following information from the internet.
Kindly verify the authenticity of the information.

The reason you're seeing a CPU limit of 1m instead of the requested 0.2m in your pod description is due to Kubernetes rounding up resource requests and limits.

Here's a breakdown of what's happening:

Requested CPU: 0.2m (200 millicores) - This is the value you specified in your pod.yaml file.
Rounding: Kubernetes rounds up resource requests and limits to the nearest whole millicore. In your case, 0.2m gets rounded up to 1m (1000 millicores).
This rounding behavior applies to both CPU requests and limits. It's important to be aware of this when setting resource constraints for your pods.

Here are some additional points to consider:

Impact: While rounding up to the nearest millicore might seem insignificant, it can affect resource allocation, especially for clusters with limited CPU resources. If you have many pods requesting small CPU amounts, rounding up could lead to oversubscription and potential performance issues.
Alternatives: There are a few ways to address this rounding issue:
Request the exact amount: If precise control is crucial, you can request the exact amount of CPU needed by your pod (e.g., 0.2m). However, keep in mind that requesting too low a value might lead to pod scheduling failures if there aren't enough resources available.
Over-request slightly: You can strategically request a slightly higher value than what your pod actually needs. This can help avoid scheduling issues due to rounding, but be mindful not to over-request significantly, as it can waste cluster resources.
By understanding Kubernetes resource rounding behavior, you can make informed decisions when setting CPU requests and limits for your pods, ensuring efficient resource utilization and optimal pod performance.
#5

0.2m does not mean 200 millicores. m suffix means millicore, therefore 0.2m is asking for 200 microcores

With CPU resources, no unit means whole cores, therefore

  • cpu: 2 - Two whole cores
  • cpu: 200m - 200 millicores, i.e. 1/5 of a core
  • cpu: 0.2 - same as 200m
  • cpu: 200u - 200 microcores, will be rounded up to 1m