ClusterFk Chaos Proxy is an unreliable HTTP proxy you can rely on; a lightweight tool designed for chaos testing of microservices.

I will let you in on a secret: everything fails eventually. Micro-services often communicate with other services via REST and HTTP. How does your micro-service cope when the services it depends on inevitably fail?

• Configure your locally running service-under-test to point to the chaos proxy and configure the chaos proxy to point to your real running dependent-destination-service.

• Switch on ClusterFk chaos proxy and configure a "chaos strategy".

• Watch the world burn 🔥 🔥 🔥

• (Optional) Do something about it.

Pull the latest image:

docker pull andymacdonald/clusterf-chaos-proxy

Then follow the steps in Running and Configuration to configure the proxy.

• Build project and create a new docker image:

mvn clean package && mvn docker:build

For running ClusterFk Chaos Proxy it is recommended you run using Docker-Compose. Define a docker-compose.yml file such as the below:

version: "3.7"
services:
  user-service-chaos-proxy:
    image: andymacdonald/clusterf-chaos-proxy
    environment:
      JAVA_OPTS: "-Dchaos.strategy=RANDOM_HAVOC -Ddestination.hostProtocolAndPort=http://localhost:8098"
    ports:
      - "8080:8080"

Then simply run (where the docker-compose.yml file is located):

docker-compose up

This will allow you to run multiple instances of ClusterFk Chaos Proxy - e.g:

version: "3.7"
services:
  user-service-chaos-proxy:
    image: andymacdonald/clusterf-chaos-proxy
    environment:
      JAVA_OPTS: "-Dchaos.strategy=RANDOM_HAVOC -Ddestination.hostProtocolAndPort=http://10.0.0.231:8098"
    ports:
      - "8080:8080"
  account-service-chaos-proxy:
    image: andymacdonald/clusterf-chaos-proxy
    environment:
      JAVA_OPTS: "-Dchaos.strategy=DELAY_RESPONSE -Ddestination.hostProtocolAndPort=http://10.0.1.150:8918"
    ports:
      - "8081:8080"

• Create a config directory containing your application.properties and take note of the directory name.

• Run application - swapping <LOCATION_OF_CONFIG> for the directory from the step before:

java -jar clusterf-chaos-proxy.jar -Dspring.config.location=<LOCATION_OF_CONFIG>/config/application.properties

Specify the port you would like to run clusterfk chaos proxy on:

server.port=8080 #8080 is the default

Use this information to configure your service-under-test with relevant config for ClusterFk chaos proxy in place of your dependent-destination-service.

Configure Chaos Proxy

Within ClusterFk chaos proxy, specify application.properties or JAVA_OPTS to point to your real destination service - e.g.:

destination.hostProtocolAndPort=http://10.0.1.150:9898

Chaos Strategies

Specify your chaos strategy:

NO_CHAOS - Request is simply passed through

DELAY_RESPONSE - Requests are delayed but successful (configurable delay)

INTERNAL_SERVER_ERROR - Requests return with 500 INTERNAL SERVER ERROR

BAD_REQUEST - Requests return with 400 BAD REQUEST

RANDOM_HAVOC - Requests generally succeed, but randomly fail with random HTTP status codes and random delays

Within your application.properties or JAVA_OPTS:

chaos.strategy=DELAY_RESPONSE

Delay Properties

If you are specifying a delayed response - you can customise the behaviour of the delay with the following properties:

chaos.strategy.delayResponse.fixedPeriod=true       # if number of seconds to delay is constant or random (default is false)
chaos.tracing.headers=false                         # If this is set to true then 2 headers will be added to the request, 
                                                     'x-clusterfk-delayed-by' and 'x-clusterfk-status-code'.
chaos.strategy.delayResponse.seconds=60             # if fixed-period delay - number of seconds to delay each requests
chaos.strategy.delayResponse.random.maxSeconds=120  # if delay is random - maximum amount of time a request can last

Configure Service Under Test

For example, you might configure your service-under-test to point to a user service which has been "clusterfk'd". If you configure your service-under-test with properties files, they might change like so:

user-service.baseUrl=http://10.0.1.150:9898/user-service

To:

user-service.baseUrl=http://localhost:8080/user-service