You are viewing documentation for Kubernetes version: v1.23

Kubernetes v1.23 documentation is no longer actively maintained. The version you are currently viewing is a static snapshot. For up-to-date documentation, see the latest version.

Resource Management for Pods and Containers

When you specify a Pod, you can optionally specify how much of each resource a container needs. The most common resources to specify are CPU and memory (RAM); there are others.

When you specify the resource request for containers in a Pod, the kube-scheduler uses this information to decide which node to place the Pod on. When you specify a resource limit for a container, the kubelet enforces those limits so that the running container is not allowed to use more of that resource than the limit you set. The kubelet also reserves at least the request amount of that system resource specifically for that container to use.

Requests and limits

If the node where a Pod is running has enough of a resource available, it's possible (and allowed) for a container to use more resource than its request for that resource specifies. However, a container is not allowed to use more than its resource limit.

For example, if you set a memory request of 256 MiB for a container, and that container is in a Pod scheduled to a Node with 8GiB of memory and no other Pods, then the container can try to use more RAM.

If you set a memory limit of 4GiB for that container, the kubelet (and container runtime) enforce the limit. The runtime prevents the container from using more than the configured resource limit. For example: when a process in the container tries to consume more than the allowed amount of memory, the system kernel terminates the process that attempted the allocation, with an out of memory (OOM) error.

Limits can be implemented either reactively (the system intervenes once it sees a violation) or by enforcement (the system prevents the container from ever exceeding the limit). Different runtimes can have different ways to implement the same restrictions.

Resource types

CPU and memory are each a resource type. A resource type has a base unit. CPU represents compute processing and is specified in units of Kubernetes CPUs. Memory is specified in units of bytes. For Linux workloads, you can specify huge page resources. Huge pages are a Linux-specific feature where the node kernel allocates blocks of memory that are much larger than the default page size.

For example, on a system where the default page size is 4KiB, you could specify a limit, hugepages-2Mi: 80Mi. If the container tries allocating over 40 2MiB huge pages (a total of 80 MiB), that allocation fails.

CPU and memory are collectively referred to as compute resources, or resources. Compute resources are measurable quantities that can be requested, allocated, and consumed. They are distinct from API resources. API resources, such as Pods and Services are objects that can be read and modified through the Kubernetes API server.

Resource requests and limits of Pod and container

For each container, you can specify resource limits and requests, including the following:

  • spec.containers[].resources.limits.cpu
  • spec.containers[].resources.limits.memory
  • spec.containers[].resources.limits.hugepages-<size>
  • spec.containers[].resources.requests.cpu
  • spec.containers[].resources.requests.memory
  • spec.containers[].resources.requests.hugepages-<size>

Although you can only specify requests and limits for individual containers, it is also useful to think about the overall resource requests and limits for a Pod. For a particular resource, a Pod resource request/limit is the sum of the resource requests/limits of that type for each container in the Pod.

Resource units in Kubernetes

CPU resource units

Limits and requests for CPU resources are measured in cpu units. In Kubernetes, 1 CPU unit is equivalent to 1 physical CPU core, or 1 virtual core, depending on whether the node is a physical host or a virtual machine running inside a physical machine.

Fractional requests are allowed. When you define a container with spec.containers[].resources.requests.cpu set to 0.5, you are requesting half as much CPU time compared to if you asked for 1.0 CPU. For CPU resource units, the quantity expression 0.1 is equivalent to the expression 100m, which can be read as "one hundred millicpu". Some people say "one hundred millicores", and this is understood to mean the same thing.

CPU resource is always specified as an absolute amount of resource, never as a relative amount. For example, 500m CPU represents the roughly same amount of computing power whether that container runs on a single-core, dual-core, or 48-core machine.

Memory resource units

Limits and requests for memory are measured in bytes. You can express memory as a plain integer or as a fixed-point number using one of these quantity suffixes: E, P, T, G, M, k. You can also use the power-of-two equivalents: Ei, Pi, Ti, Gi, Mi, Ki. For example, the following represent roughly the same value:

128974848, 129e6, 129M,  128974848000m, 123Mi

Pay attention to the case of the suffixes. If you request 400m of memory, this is a request for 0.4 bytes. Someone who types that probably meant to ask for 400 mebibytes (400Mi) or 400 megabytes (400M).

Container resources example

The following Pod has two containers. Both containers are defined with a request for 0.25 CPU and 64MiB (226 bytes) of memory. Each container has a limit of 0.5 CPU and 128MiB of memory. You can say the Pod has a request of 0.5 CPU and 128 MiB of memory, and a limit of 1 CPU and 256MiB of memory.

apiVersion: v1
kind: Pod
  name: frontend
  - name: app
        memory: "64Mi"
        cpu: "250m"
        memory: "128Mi"
        cpu: "500m"
  - name: log-aggregator
        memory: "64Mi"
        cpu: "250m"
        memory: "128Mi"
        cpu: "500m"

How Pods with resource requests are scheduled

When you create a Pod, the Kubernetes scheduler selects a node for the Pod to run on. Each node has a maximum capacity for each of the resource types: the amount of CPU and memory it can provide for Pods. The scheduler ensures that, for each resource type, the sum of the resource requests of the scheduled containers is less than the capacity of the node. Note that although actual memory or CPU resource usage on nodes is very low, the scheduler still refuses to place a Pod on a node if the capacity check fails. This protects against a resource shortage on a node when resource usage later increases, for example, during a daily peak in request rate.

How Kubernetes applies resource requests and limits

When the kubelet starts a container as part of a Pod, the kubelet passes that container's requests and limits for memory and CPU to the container runtime.

On Linux, the container runtime typically configures kernel cgroups that apply and enforce the limits you defined.

  • The CPU limit defines a hard ceiling on how much CPU time that the container can use. During each scheduling interval (time slice), the Linux kernel checks to see if this limit is exceeded; if so, the kernel waits before allowing that cgroup to resume execution.
  • The CPU request typically defines a weighting. If several different containers (cgroups) want to run on a contended system, workloads with larger CPU requests are allocated more CPU time than workloads with small requests.
  • The memory request is mainly used during (Kubernetes) Pod scheduling. On a node that uses cgroups v2, the container runtime might use the memory request as a hint to set memory.min and memory.low.
  • The memory limit defines a memory limit for that cgroup. If the container tries to allocate more memory than this limit, the Linux kernel out-of-memory subsystem activates and, typically, intervenes by stopping one of the processes in the container that tried to allocate memory. If that process is the container's PID 1, and the container is marked as restartable, Kubernetes restarts the container.
  • The memory limit for the Pod or container can also apply to pages in memory backed volumes, such as an emptyDir. The kubelet tracks tmpfs emptyDir volumes as container memory use, rather than as local ephemeral storage.

If a container exceeds its memory request and the node that it runs on becomes short of memory overall, it is likely that the Pod the container belongs to will be evicted.

A container might or might not be allowed to exceed its CPU limit for extended periods of time. However, container runtimes don't terminate Pods or containers for excessive CPU usage.

To determine whether a container cannot be scheduled or is being killed due to resource limits, see the Troubleshooting section.

Monitoring compute & memory resource usage

The kubelet reports the resource usage of a Pod as part of the Pod status.

If optional tools for monitoring are available in your cluster, then Pod resource usage can be retrieved either from the Metrics API directly or from your monitoring tools.

Local ephemeral storage

FEATURE STATE: Kubernetes v1.10 [beta]

Nodes have local ephemeral storage, backed by locally-attached writeable devices or, sometimes, by RAM. "Ephemeral" means that there is no long-term guarantee about durability.

Pods use ephemeral local storage for scratch space, caching, and for logs. The kubelet can provide scratch space to Pods using local ephemeral storage to mount emptyDir volumes into containers.

The kubelet also uses this kind of storage to hold node-level container logs, container images, and the writable layers of running containers.

As a beta feature, Kubernetes lets you track, reserve and limit the amount of ephemeral local storage a Pod can consume.

Configurations for local ephemeral storage

Kubernetes supports two ways to configure local ephemeral storage on a node:

In this configuration, you place all different kinds of ephemeral local data (emptyDir volumes, writeable layers, container images, logs) into one filesystem. The most effective way to configure the kubelet means dedicating this filesystem to Kubernetes (kubelet) data.

The kubelet also writes node-level container logs and treats these similarly to ephemeral local storage.

The kubelet writes logs to files inside its configured log directory (/var/log by default); and has a base directory for other locally stored data (/var/lib/kubelet by default).

Typically, both /var/lib/kubelet and /var/log are on the system root filesystem, and the kubelet is designed with that layout in mind.

Your node can have as many other filesystems, not used for Kubernetes, as you like.

You have a filesystem on the node that you're using for ephemeral data that comes from running Pods: logs, and emptyDir volumes. You can use this filesystem for other data (for example: system logs not related to Kubernetes); it can even be the root filesystem.

The kubelet also writes node-level container logs into the first filesystem, and treats these similarly to ephemeral local storage.

You also use a separate filesystem, backed by a different logical storage device. In this configuration, the directory where you tell the kubelet to place container image layers and writeable layers is on this second filesystem.

The first filesystem does not hold any image layers or writeable layers.

Your node can have as many other filesystems, not used for Kubernetes, as you like.

The kubelet can measure how much local storage it is using. It does this provided that:

  • the LocalStorageCapacityIsolation feature gate is enabled (the feature is on by default), and
  • you have set up the node using one of the supported configurations for local ephemeral storage.

If you have a different configuration, then the kubelet does not apply resource limits for ephemeral local storage.

Setting requests and limits for local ephemeral storage

You can specify ephemeral-storage for managing local ephemeral storage. Each container of a Pod can specify either or both of the following:

  • spec.containers[].resources.limits.ephemeral-storage
  • spec.containers[].resources.requests.ephemeral-storage

Limits and requests for ephemeral-storage are measured in byte quantities. You can express storage as a plain integer or as a fixed-point number using one of these suffixes: E, P, T, G, M, K. You can also use the power-of-two equivalents: Ei, Pi, Ti, Gi, Mi, Ki. For example, the following quantities all represent roughly the same value:

  • 128974848
  • 129e6
  • 129M
  • 123Mi

In the following example, the Pod has two containers. Each container has a request of 2GiB of local ephemeral storage. Each container has a limit of 4GiB of local ephemeral storage. Therefore, the Pod has a request of 4GiB of local ephemeral storage, and a limit of 8GiB of local ephemeral storage.

apiVersion: v1
kind: Pod
  name: frontend
  - name: app
        ephemeral-storage: "2Gi"
        ephemeral-storage: "4Gi"
    - name: ephemeral
      mountPath: "/tmp"
  - name: log-aggregator
        ephemeral-storage: "2Gi"
        ephemeral-storage: "4Gi"
    - name: ephemeral
      mountPath: "/tmp"
    - name: ephemeral
      emptyDir: {}

How Pods with ephemeral-storage requests are scheduled

When you create a Pod, the Kubernetes scheduler selects a node for the Pod to run on. Each node has a maximum amount of local ephemeral storage it can provide for Pods. For more information, see Node Allocatable.

The scheduler ensures that the sum of the resource requests of the scheduled containers is less than the capacity of the node.

Ephemeral storage consumption management

If the kubelet is managing local ephemeral storage as a resource, then the kubelet measures storage use in:

  • emptyDir volumes, except tmpfs emptyDir volumes
  • directories holding node-level logs
  • writeable container layers

If a Pod is using more ephemeral storage than you allow it to, the kubelet sets an eviction signal that triggers Pod eviction.

For container-level isolation, if a container's writable layer and log usage exceeds its storage limit, the kubelet marks the Pod for eviction.

For pod-level isolation the kubelet works out an overall Pod storage limit by summing the limits for the containers in that Pod. In this case, if the sum of the local ephemeral storage usage from all containers and also the Pod's emptyDir volumes exceeds the overall Pod storage limit, then the kubelet also marks the Pod for eviction.

The kubelet supports different ways to measure Pod storage use:

The kubelet performs regular, scheduled checks that scan each emptyDir volume, container log directory, and writeable container layer.

The scan measures how much space is used.

FEATURE STATE: Kubernetes v1.15 [alpha]

Project quotas are an operating-system level feature for managing storage use on filesystems. With Kubernetes, you can enable project quotas for monitoring storage use. Make sure that the filesystem backing the emptyDir volumes, on the node, provides project quota support. For example, XFS and ext4fs offer project quotas.

Kubernetes uses project IDs starting from 1048576. The IDs in use are registered in /etc/projects and /etc/projid. If project IDs in this range are used for other purposes on the system, those project IDs must be registered in /etc/projects and /etc/projid so that Kubernetes does not use them.

Quotas are faster and more accurate than directory scanning. When a directory is assigned to a project, all files created under a directory are created in that project, and the kernel merely has to keep track of how many blocks are in use by files in that project. If a file is created and deleted, but has an open file descriptor, it continues to consume space. Quota tracking records that space accurately whereas directory scans overlook the storage used by deleted files.

If you want to use project quotas, you should:

  • Enable the LocalStorageCapacityIsolationFSQuotaMonitoring=true feature gate using the featureGates field in the kubelet configuration or the --feature-gates command line flag.

  • Ensure that the root filesystem (or optional runtime filesystem) has project quotas enabled. All XFS filesystems support project quotas. For ext4 filesystems, you need to enable the project quota tracking feature while the filesystem is not mounted.

    # For ext4, with /dev/block-device not mounted
    sudo tune2fs -O project -Q prjquota /dev/block-device
  • Ensure that the root filesystem (or optional runtime filesystem) is mounted with project quotas enabled. For both XFS and ext4fs, the mount option is named prjquota.

Extended resources

Extended resources are fully-qualified resource names outside the domain. They allow cluster operators to advertise and users to consume the non-Kubernetes-built-in resources.

There are two steps required to use Extended Resources. First, the cluster operator must advertise an Extended Resource. Second, users must request the Extended Resource in Pods.

Managing extended resources

Node-level extended resources

Node-level extended resources are tied to nodes.

Device plugin managed resources

See Device Plugin for how to advertise device plugin managed resources on each node.

Other resources

To advertise a new node-level extended resource, the cluster operator can submit a PATCH HTTP request to the API server to specify the available quantity in the status.capacity for a node in the cluster. After this operation, the node's status.capacity will include a new resource. The status.allocatable field is updated automatically with the new resource asynchronously by the kubelet.

Because the scheduler uses the node's status.allocatable value when evaluating Pod fitness, the scheduler only takes account of the new value after that asynchronous update. There may be a short delay between patching the node capacity with a new resource and the time when the first Pod that requests the resource can be scheduled on that node.


Here is an example showing how to use curl to form an HTTP request that advertises five "" resources on node k8s-node-1 whose master is k8s-master.

curl --header "Content-Type: application/json-patch+json" \
--request PATCH \
--data '[{"op": "add", "path": "/status/capacity/", "value": "5"}]' \

Cluster-level extended resources

Cluster-level extended resources are not tied to nodes. They are usually managed by scheduler extenders, which handle the resource consumption and resource quota.

You can specify the extended resources that are handled by scheduler extenders in scheduler configuration


The following configuration for a scheduler policy indicates that the cluster-level extended resource "" is handled by the scheduler extender.

  • The scheduler sends a Pod to the scheduler extender only if the Pod requests "".
  • The ignoredByScheduler field specifies that the scheduler does not check the "" resource in its PodFitsResources predicate.
  "kind": "Policy",
  "apiVersion": "v1",
  "extenders": [
      "bindVerb": "bind",
      "managedResources": [
          "name": "",
          "ignoredByScheduler": true

Consuming extended resources

Users can consume extended resources in Pod specs like CPU and memory. The scheduler takes care of the resource accounting so that no more than the available amount is simultaneously allocated to Pods.

The API server restricts quantities of extended resources to whole numbers. Examples of valid quantities are 3, 3000m and 3Ki. Examples of invalid quantities are 0.5 and 1500m.

To consume an extended resource in a Pod, include the resource name as a key in the spec.containers[].resources.limits map in the container spec.

A Pod is scheduled only if all of the resource requests are satisfied, including CPU, memory and any extended resources. The Pod remains in the PENDING state as long as the resource request cannot be satisfied.


The Pod below requests 2 CPUs and 1 "" (an extended resource).

apiVersion: v1
kind: Pod
  name: my-pod
  - name: my-container
    image: myimage
        cpu: 2 1
      limits: 1

PID limiting

Process ID (PID) limits allow for the configuration of a kubelet to limit the number of PIDs that a given Pod can consume. See PID Limiting for information.


My Pods are pending with event message FailedScheduling

If the scheduler cannot find any node where a Pod can fit, the Pod remains unscheduled until a place can be found. An Event is produced each time the scheduler fails to find a place for the Pod. You can use kubectl to view the events for a Pod; for example:

kubectl describe pod frontend | grep -A 9999999999 Events
  Type     Reason            Age   From               Message
  ----     ------            ----  ----               -------
  Warning  FailedScheduling  23s   default-scheduler  0/42 nodes available: insufficient cpu

In the preceding example, the Pod named "frontend" fails to be scheduled due to insufficient CPU resource on any node. Similar error messages can also suggest failure due to insufficient memory (PodExceedsFreeMemory). In general, if a Pod is pending with a message of this type, there are several things to try:

  • Add more nodes to the cluster.
  • Terminate unneeded Pods to make room for pending Pods.
  • Check that the Pod is not larger than all the nodes. For example, if all the nodes have a capacity of cpu: 1, then a Pod with a request of cpu: 1.1 will never be scheduled.
  • Check for node taints. If most of your nodes are tainted, and the new Pod does not tolerate that taint, the scheduler only considers placements onto the remaining nodes that don't have that taint.

You can check node capacities and amounts allocated with the kubectl describe nodes command. For example:

kubectl describe nodes e2e-test-node-pool-4lw4
Name:            e2e-test-node-pool-4lw4
[ ... lines removed for clarity ...]
 cpu:                               2
 memory:                            7679792Ki
 pods:                              110
 cpu:                               1800m
 memory:                            7474992Ki
 pods:                              110
[ ... lines removed for clarity ...]
Non-terminated Pods:        (5 in total)
  Namespace    Name                                  CPU Requests  CPU Limits  Memory Requests  Memory Limits
  ---------    ----                                  ------------  ----------  ---------------  -------------
  kube-system  fluentd-gcp-v1.38-28bv1               100m (5%)     0 (0%)      200Mi (2%)       200Mi (2%)
  kube-system  kube-dns-3297075139-61lj3             260m (13%)    0 (0%)      100Mi (1%)       170Mi (2%)
  kube-system  kube-proxy-e2e-test-...               100m (5%)     0 (0%)      0 (0%)           0 (0%)
  kube-system  monitoring-influxdb-grafana-v4-z1m12  200m (10%)    200m (10%)  600Mi (8%)       600Mi (8%)
  kube-system  node-problem-detector-v0.1-fj7m3      20m (1%)      200m (10%)  20Mi (0%)        100Mi (1%)
Allocated resources:
  (Total limits may be over 100 percent, i.e., overcommitted.)
  CPU Requests    CPU Limits    Memory Requests    Memory Limits
  ------------    ----------    ---------------    -------------
  680m (34%)      400m (20%)    920Mi (11%)        1070Mi (13%)

In the preceding output, you can see that if a Pod requests more than 1.120 CPUs or more than 6.23Gi of memory, that Pod will not fit on the node.

By looking at the “Pods” section, you can see which Pods are taking up space on the node.

The amount of resources available to Pods is less than the node capacity because system daemons use a portion of the available resources. Within the Kubernetes API, each Node has a .status.allocatable field (see NodeStatus for details).

The .status.allocatable field describes the amount of resources that are available to Pods on that node (for example: 15 virtual CPUs and 7538 MiB of memory). For more information on node allocatable resources in Kubernetes, see Reserve Compute Resources for System Daemons.

You can configure resource quotas to limit the total amount of resources that a namespace can consume. Kubernetes enforces quotas for objects in particular namespace when there is a ResourceQuota in that namespace. For example, if you assign specific namespaces to different teams, you can add ResourceQuotas into those namespaces. Setting resource quotas helps to prevent one team from using so much of any resource that this over-use affects other teams.

You should also consider what access you grant to that namespace: full write access to a namespace allows someone with that access to remove any resource, including a configured ResourceQuota.

My container is terminated

Your container might get terminated because it is resource-starved. To check whether a container is being killed because it is hitting a resource limit, call kubectl describe pod on the Pod of interest:

kubectl describe pod simmemleak-hra99

The output is similar to:

Name:                           simmemleak-hra99
Namespace:                      default
Image(s):                       saadali/simmemleak
Node:                           kubernetes-node-tf0f/
Labels:                         name=simmemleak
Status:                         Running
    Image:  saadali/simmemleak:latest
      cpu:          100m
      memory:       50Mi
    State:          Running
      Started:      Tue, 07 Jul 2019 12:54:41 -0700
    Last State:     Terminated
      Reason:       OOMKilled
      Exit Code:    137
      Started:      Fri, 07 Jul 2019 12:54:30 -0700
      Finished:     Fri, 07 Jul 2019 12:54:33 -0700
    Ready:          False
    Restart Count:  5
  Type      Status
  Ready     False
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Normal  Scheduled  42s   default-scheduler  Successfully assigned simmemleak-hra99 to kubernetes-node-tf0f
  Normal  Pulled     41s   kubelet            Container image "saadali/simmemleak:latest" already present on machine
  Normal  Created    41s   kubelet            Created container simmemleak
  Normal  Started    40s   kubelet            Started container simmemleak
  Normal  Killing    32s   kubelet            Killing container with id ead3fb35-5cf5-44ed-9ae1-488115be66c6: Need to kill Pod

In the preceding example, the Restart Count: 5 indicates that the simmemleak container in the Pod was terminated and restarted five times (so far). The OOMKilled reason shows that the container tried to use more memory than its limit.

Your next step might be to check the application code for a memory leak. If you find that the application is behaving how you expect, consider setting a higher memory limit (and possibly request) for that container.

What's next

Last modified April 30, 2022 at 9:21 AM PST: [en] modify link about debug (e6276724bb)