kubernetesschedulingPractitioner

Mastering Resource Management for Kubernetes Pods

5 min read Kubernetes DocsApr 22, 2026Reviewed for accuracy

Practitioner — Hands-on experience recommended

Resource management in Kubernetes is essential for maintaining application performance and stability. By defining resource requests and limits for your Pods and containers, you can control how much CPU and memory each application can consume. This prevents resource contention and ensures that critical applications have the resources they need to function correctly.

When you specify a resource request for a container, the kube-scheduler uses this information to decide which node to place the Pod on. Resource limits are enforced by the kubelet, preventing containers from exceeding their allocated resources. If a node has enough resources available, a container can use more than its request specifies, but it will be throttled if it exceeds its limit. CPU limits are enforced through throttling, while memory limits are managed by the kernel, which may kill processes if they exceed their memory allocation. You can define these parameters in your Pod specifications using fields like spec.containers[].resources.requests.cpu and spec.containers[].resources.limits.memory.

In production, it's crucial to understand the implications of these settings. For instance, if you set a limit without a corresponding request, Kubernetes will use the limit as the request value, which can lead to unexpected scheduling behavior. Additionally, be cautious with CPU specifications; Kubernetes does not allow precision finer than 0.001 CPU, so using milliCPU is recommended for values less than one. Keep an eye on the alpha feature MemoryQoS, which aims to improve memory limit enforcement, but be aware of its current limitations and potential issues.

Key takeaways

→Define resource requests to guide the kube-scheduler in Pod placement.
→Set resource limits to prevent containers from consuming excessive resources.
→Use milliCPU for CPU specifications to avoid precision issues.
→Be cautious with limits without corresponding requests to prevent unexpected behavior.
→Monitor the MemoryQoS feature for potential improvements in memory management.

Why it matters

Proper resource management directly impacts application performance and stability in production environments. Misconfigurations can lead to resource contention, degraded performance, or even application crashes.

Code examples

YAML

1---
2apiVersion: v1
3kind: Pod
4metadata:
5  name: frontend
6spec:
7  containers:
8  - name: app
9    image: images.my-company.example/app:v4
10    resources:
11      requests:
12        memory: "64Mi"
13        cpu: "250m"
14      limits:
15        memory: "128Mi"
16        cpu: "500m"
17  - name: log-aggregator
18    image: images.my-company.example/log-aggregator:v6
19    resources:
20      requests:
21        memory: "64Mi"
22        cpu: "250m"
23      limits:
24        memory: "128Mi"
25        cpu: "500m"

YAML

1apiVersion: v1
2kind: Pod
3metadata:
4  name: pod-resources-demo
5  namespace: pod-resources-example
6spec:
7  resources:
8    limits:
9      cpu: "1"
10      memory: "200Mi"
11    requests:
12      cpu: "1"
13      memory: "100Mi"
14  containers:
15  - name: pod-resources-demo-ctr-1
16    image: nginx
17    resources:
18      limits:
19        cpu: "0.5"
20        memory: "100Mi"
21      requests:
22        cpu: "0.5"
23        memory: "50Mi"
24  - name: pod-resources-demo-ctr-2
25    image: fedora
26    command:
27    - sleep
28    - -inf