Simulate Stress Scenarios

StressChaos Introduction

Chaos Mesh provides StressChaos experiments to simulate stress scenarios inside containers. This document describes how to create StressChaos experiments and how to prepare the corresponding configuration file.

You can create experiments using either Chaos Dashboard or the YAML configuration files.

Create experiments using Chaos Dashboard

1. Open Chaos Dashboard, and click NEW EXPERIMENT on the page to create a new experiment:

   

   Create Experiment

2. In the Choose a Target area, choose STRESS TEST and fill out the experiment information. For specific configuration fields, refer to the description in [Configuration Description](#fields description).

   

   StressChaos Experiment

3. Fill out the experiment information, and specify the experiment scope and the scheduled experiment duration:

   

   Experiment Information

4. Submit the experiment information.

Create experiments using the YAML file

1. Write the experiment configuration to the YAML configuration file. In the following example, the memory-stress.yaml file is used.

   apiVersion: chaos-mesh.org/v1alpha1
   kind: StressChaos
   metadata:
     name: memory-stress-example
     namespace: chaos-mesh
   spec:
     mode: one
     selector:
       labelSelectors:
         'app': 'app1'
     stressors:
       memory:
         workers: 4
         size: '256MB'

   This experiment configuration will create a process in the selected container, continuously allocate and read and write in memory, occupying up to 256MB of memory.

2. After the configuration file is prepared, use kubectl to create an experiment:

   kubectl apply -f memory-stress.yaml

Field description

The fields in the YAML configuration file are described in the following table:

Parameter	Type	Description	Default value	Required	Example
duration	string	Specifies the duration of the experiment.	None	Yes	30s
stressors	Stressors	Specifies the stress of CPU or memory	None	No
stressngStressors	string	Specifies the stres-ng parameter to reach richer stress injection	None	No	--clone 2
mode	string	Specifies the mode of the experiment. The mode options include one (selecting a random Pod), all (selecting all eligible Pods), fixed (selecting a specified number of eligible Pods), fixed-percent (selecting a specified percentage of Pods from the eligible Pods), and random-max-percent (selecting the maximum percentage of Pods from the eligible Pods).	None	Yes	one
value	string	Provides a parameter for the mode configuration, depending on mode.For example, when mode is set to fixed-percent, value specifies the percentage of Pods.	None	No	1
containerNames	[]string	Specifies the name of the container into which the fault is injected.	None	No	["nginx"]
selector	struct	Specifies the target Pod. For details, refer to Define the Scope of Chaos Experiments.	None	Yes

Stressors

Parameter	Type	Description	Default value	Required	Example
memory	MemoryStressor	Specifies the memory stress	None	No
cpu	CPUStressor	Specifies the CPU stress	None	No

MemoryStressor

Parameter	Type	Description	Default value	Required	Example
workers	int	Specifies the number of threads that apply memory stress		No	1
size	string	Specifies the memory size to be occupied or a percentage of the total memory size. The final sum of the occupied memory size is size.		No	256MB / 25%
time	string	Specifies the time to reach the memory size. The growth model is a linear model.		No	10min
oomScoreAdj	int	Specifies the oom_score_adj of the stress process.		No	-1000

note

To avoid the high CPU load caused by the read and write pressure from stress-ng, Chaos Mesh uses memStress to simulate memory stress. This is because memStress simulates memory stress by consuming actual memory instead of applying the read and write pressure to memory.

CPUStressor

Parameter	Type	Description	Default value	Required	Example
workers	int	Specifies the number of threads that apply CPU stress		Yes	1
load	int	Specifies the percentage of CPU occupied. 0 means that no additional CPU is added, and 100 refers to full load. The final sum of CPU load is workers * load.		No	50