litmus/pod-cpu-hog-exec.md at 2.2.0

9.7 KiB

Raw Permalink Blame History

Introduction

This experiment consumes the CPU resources of the application container
It simulates conditions where app pods experience CPU spikes either due to expected/undesired processes thereby testing how the overall application stack behaves when this occurs.

!!! tip "Scenario: Stress the CPU"

Uses

??? info "View the uses of the experiment" Disk Pressure or CPU hogs is another very common and frequent scenario we find in kubernetes applications that can result in the eviction of the application replica and impact its delivery. Such scenarios that can still occur despite whatever availability aids K8s provides. These problems are generally referred to as "Noisy Neighbour" problems.

Injecting a rogue process into a target container, we starve the main microservice process (typically pid 1) of the resources allocated to it (where limits are defined) causing slowness in application traffic or in other cases unrestrained use can cause node to exhaust resources leading to eviction of all pods.So this category of chaos experiment helps to build the immunity on the application undergoing any such stress scenario

Prerequisites

??? info "Verify the prerequisites" - Ensure that Kubernetes Version > 1.16 - Ensure that the Litmus Chaos Operator is running by executing kubectl get pods in operator namespace (typically, litmus).If not, install from here - Ensure that the pod-cpu-hog-exec experiment resource is available in the cluster by executing kubectl get chaosexperiments in the desired namespace. If not, install from here

Default Validations

??? info "View the default validations" The application pods should be in running state before and after chaos injection.

Minimal RBAC configuration example (optional)

!!! tip "NOTE"
If you are using this experiment as part of a litmus workflow scheduled constructed & executed from chaos-center, then you may be making use of the litmus-admin RBAC, which is pre installed in the cluster as part of the agent setup.

??? note "View the Minimal RBAC permissions"

    [embedmd]:# (https://raw.githubusercontent.com/litmuschaos/chaos-charts/master/charts/generic/pod-cpu-hog-exec/rbac.yaml yaml)
    ```yaml
    ---
    apiVersion: v1
    kind: ServiceAccount
    metadata:
      name: pod-cpu-hog-exec-sa
      namespace: default
      labels:
        name: pod-cpu-hog-exec-sa
        app.kubernetes.io/part-of: litmus
    ---
    apiVersion: rbac.authorization.k8s.io/v1
    kind: Role
    metadata:
      name: pod-cpu-hog-exec-sa
      namespace: default
      labels:
        name: pod-cpu-hog-exec-sa
        app.kubernetes.io/part-of: litmus
    rules:
    - apiGroups: [""]
      resources: ["pods","events"]
      verbs: ["create","list","get","patch","update","delete","deletecollection"]
    - apiGroups: [""]
      resources: ["pods/exec","pods/log","replicationcontrollers"]
      verbs: ["create","list","get"]
    - apiGroups: ["batch"]
      resources: ["jobs"]
      verbs: ["create","list","get","delete","deletecollection"]
    - apiGroups: ["apps"]
      resources: ["deployments","statefulsets","daemonsets","replicasets"]
      verbs: ["list","get"]
    - apiGroups: ["apps.openshift.io"]
      resources: ["deploymentconfigs"]
      verbs: ["list","get"]
    - apiGroups: ["argoproj.io"]
      resources: ["rollouts"]
      verbs: ["list","get"]
    - apiGroups: ["litmuschaos.io"]
      resources: ["chaosengines","chaosexperiments","chaosresults"]
      verbs: ["create","list","get","patch","update"]
    ---
    apiVersion: rbac.authorization.k8s.io/v1
    kind: RoleBinding
    metadata:
      name: pod-cpu-hog-exec-sa
      namespace: default
      labels:
        name: pod-cpu-hog-exec-sa
        app.kubernetes.io/part-of: litmus
    roleRef:
      apiGroup: rbac.authorization.k8s.io
      kind: Role
      name: pod-cpu-hog-exec-sa
    subjects:
    - kind: ServiceAccount
      name: pod-cpu-hog-exec-sa
      namespace: default
    ```
    Use this sample RBAC manifest to create a chaosServiceAccount in the desired (app) namespace. This example consists of the minimum necessary role permissions to execute the experiment.

Experiment tunables

??? info "check the experiment tunables"

Optional Fields

<table>
  <tr>
    <th> Variables </th>
    <th> Description </th>
    <th> Notes </th>
  </tr>
  <tr>
    <td> CPU_CORES </td>
    <td> Number of the cpu cores subjected to CPU stress  </td>
    <td> Default to 1 </td>
  </tr>
  <tr>
    <td> TOTAL_CHAOS_DURATION </td>
    <td> The time duration for chaos insertion (seconds)  </td>
    <td> Default to 60s </td>
  </tr>
  <tr>
    <td> LIB  </td>
    <td> The chaos lib used to inject the chaos. Available libs are <code>litmus</code></td>
    <td> Default to <code>litmus</code> </td>
  </tr>
  <tr>
    <td> TARGET_PODS </td>
    <td> Comma separated list of application pod name subjected to pod cpu hog chaos</td>
    <td> If not provided, it will select target pods randomly based on provided appLabels</td>
  </tr> 
  <tr> 
    <td> TARGET_CONTAINER </td>
    <td> Name of the target container under chaos </td>
    <td> If not provided, it will select the first container of the target pod </td>
  </tr> 
  <tr>
    <td> PODS_AFFECTED_PERC </td>
    <td> The Percentage of total pods to target  </td>
    <td> Defaults to 0 (corresponds to 1 replica), provide numeric value only </td>
  </tr>
   <tr>
    <td> CHAOS_INJECT_COMMAND </td>
    <td> The command to inject the cpu chaos  </td>
    <td> Default to <code>md5sum /dev/zero</code> </td>
  </tr>
  <tr>
    <td> CHAOS_KILL_COMMAND </td>
    <td> The command to kill the chaos process</td>
    <td> Default to <code>kill $(find /proc -name exe -lname '*/md5sum' 2>&1 | grep -v 'Permission denied' | awk -F/ '{print $(NF-1)}')</code>. Another useful one that generally works (in case the default doesn't) is <code>kill -9 $(ps afx | grep \"[md5sum] /dev/zero\" | awk '{print$1}' | tr '\n' ' ')</code>. In case neither works, please check whether the target pod's base image offers a shell. If yes, identify appropriate shell command to kill the chaos process </td>
  </tr>
  <tr>
    <td> RAMP_TIME </td>
    <td> Period to wait before injection of chaos in sec </td>
    <td> </td>
  </tr>
  <tr>
    <td> SEQUENCE </td>
    <td> It defines sequence of chaos execution for multiple target pods </td>
    <td> Default value: parallel. Supported: serial, parallel </td>
  </tr>
</table>

Experiment Examples

Common and Pod specific tunables

Refer the common attributes and Pod specific tunable to tune the common tunables for all experiments and pod specific tunables.

CPU Cores

It stresses the CPU_CORE cpu cores of the targeted pod for the TOTAL_CHAOS_DURATION duration.

Use the following example to tune this:

# cpu cores for the stress
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: pod-cpu-hog-exec-sa
  experiments:
  - name: pod-cpu-hog-exec
    spec:
      components:
        env:
        # cpu cores for stress
        - name: CPU_CORES
          value: '1'
        - name: TOTAL_CHAOS_DURATION
          value: '60'

Chaos Inject and Kill Commands

It defines the CHAOS_INJECT_COMMAND and CHAOS_KILL_COMMAND ENV to set the chaos inject and chaos kill commands respectively. Default values of commands:

CHAOS_INJECT_COMMAND: "md5sum /dev/zero"
CHAOS_KILL_COMMAND: "kill (find /proc -name exe -lname '*/md5sum' 2>&1 | grep -v 'Permission denied' | awk -F/ '{print (NF-1)}')"