Simulate Stress Scenarios

This document describes how to use Chaosd to simulate stress scenarios. This feature generates CPU or memory stress on the host using stress-ng. You can create stress experiments either in command-line or service mode.

Create stress experiments using the command-line mode

This section describes how to create stress experiments using the command-line mode.

Before creating stress experiments, you can run the following command to view the stress experiment types supported by Chaosd:

chaosd attack stress --help

The result is as follows:

Stress attack related commands
Usage:
  chaosd attack stress [command]
Available Commands:
  cpu         continuously stress CPU out
  mem         continuously stress virtual memory out
Flags:
  -h, --help   help for stress
Global Flags:
      --log-level string   the log level of chaosd, the value can be 'debug', 'info', 'warn' and 'error'
Use "chaosd attack stress [command] --help" for more information about a command.

Currently, Chaosd supports creating CPU stress experiments and memory stress experiments.

Simulate CPU stress using the command-line mode

Command for simulating CPU stress

To view the configuration items supported by the CPU stress simulation, run the following command:

chaosd attack stress cpu --help

The result is as follows:

continuously stress CPU out
Usage:
  chaosd attack stress cpu [options] [flags]
Flags:
  -h, --help              help for cpu
  -l, --load int          Load specifies P percent loading per CPU worker. 0 is effectively a sleep (no load) and 100 is full loading. (default 10)
  -o, --options strings   extend stress-ng options.
  -w, --workers int       Workers specifies N workers to apply the stressor. (default 1)
Global Flags:
      --log-level string   the log level of chaosd, the value can be 'debug', 'info', 'warn' and 'error'

Configuration description for simulating CPU stress

Example for simulating CPU stress

chaosd attack stress cpu --workers 2 --load 10

The result is as follows:

[2021/05/12 03:38:33.698 +00:00] [INFO] [stress.go:66] ["stressors normalize"] [arguments=" --cpu 2 --cpu-load 10"]
[2021/05/12 03:38:33.702 +00:00] [INFO] [stress.go:82] ["Start stress-ng process successfully"] [command="/usr/bin/stress-ng --cpu 2 --cpu-load 10"] [Pid=27483]
Attack stress cpu successfully, uid: 4f33b2d4-aee6-43ca-9c43-0f12867e5c9c

Simulating memory stress using the command-line mode

Command for simulating memory stress

To view the configuration items supported by the memory stress simulation, run the following command:

chaosd attack stress mem --help

The result is as follows:

continuously stress virtual memory out
Usage:
  chaosd attack stress mem [options] [flags]
Flags:
  -h, --help              help for mem
  -o, --options strings   extend stress-ng options.
  -s, --size string       Size specifies N bytes consumed per vm worker, default is the total available memory. One can specify the size as % of total available memory or in units of B, KB/KiB, MB/MiB, GB/GiB, TB/TiB..
Global Flags:
      --log-level string   the log level of chaosd, the value can be 'debug', 'info', 'warn' and 'error'

Configuration description for simulating memory stress

Example for simulating memory stress

chaosd attack stress mem --workers 2 --size 100M

The result is as follows:

[2021/05/12 03:37:19.643 +00:00] [INFO] [stress.go:66] ["stressors normalize"] [arguments=" --vm 2 --vm-keep --vm-bytes 100000000"]
[2021/05/12 03:37:19.654 +00:00] [INFO] [stress.go:82] ["Start stress-ng process successfully"] [command="/usr/bin/stress-ng --vm 2 --vm-keep --vm-bytes 100000000"] [Pid=26799]
Attack stress mem successfully, uid: c2bff2f5-3aac-4ace-b7a6-322946ae6f13

When running the experiment, you need to save the uid information of the experiment. When a stress simulation is not needed, you can use recover to terminate the uid-related experiment::

chaosd recover c2bff2f5-3aac-4ace-b7a6-322946ae6f13

The result is as follows:

Recover c2bff2f5-3aac-4ace-b7a6-322946ae6f13 successfully

Create stress experiments using the service mode

To create experiments using the service mode, follow the instructions below:

1. Run Chaosd in service mode:

   chaosd server --port 31767

2. Send a POST HTTP request to the /api/attack/{uid} path of Chaosd service.

   curl -X POST 172.16.112.130:31767/api/attack/jvm -H "Content-Type:application/json" -d '{fault-configuration}'

   For the fault-configuration part in the above command, you need to configure it according to the fault types. For the corresponding parameters, refer to the parameters and examples of each fault type in the following sections.

note

When running an experiment, remember to save the UID information of the experiment. When you want to kill the experiment corresponding to the UID, you need to send a DELETE HTTP request to the /api/attack/{uid} path of Chaosd service.

Simulate CPU stress using the service mode

Parameters for simulating CPU stress

Example for simulating CPU stress using the service mode

curl -X POST 172.16.112.130:31767/api/attack/stress -H "Content-Type:application/json" -d '{"load":10, "action":"cpu","workers":1}'

The result is as follows:

{"status":200,"message":"attack successfully","uid":"c3c519bf-819a-4a7b-97fb-e3d0814481fa"}

Simulate memory stress using the service mode

Parameters for simulating memory stress

Example for simulating memory stress using the service mode

curl -X POST 172.16.112.130:31767/api/attack/stress -H "Content-Type:application/json" -d '{"size":"100M", "action":"mem"}'

The result is as follows:

{"status":200,"message":"attack successfully","uid":"a551206c-960d-4ac5-9056-518e512d4d0d"}