website/simulate-http-chaos-on-kubernetes.md at master

14 KiB

Raw Permalink Blame History

title
Simulate HTTP Faults

This document describes how to simulate HTTP faults by creating HTTPChaos experiments in Chaos Mesh.

HTTPChaos introduction

HTTPChaos is a fault type provided by Chaos Mesh. By creating HTTPChaos experiments, you can simulate the fault scenarios during the HTTP request and response processing. Currently, HTTPChaos supports simulating the following fault types:

abort: interrupts the connection
delay: injects latency into the request or response
replace: replaces part of content in HTTP request or response messages
patch: adds additional content to HTTP request or response messages

HTTPChaos supports combinations of different fault types. If you have configured multiple HTTP fault types at the same time when creating HTTPChaos experiments, the order set to inject the faults when the experiments start running is abort -> delay -> replace -> patch. When the abort fault cause short circuits, the connection will be directly interrupted.

For the detailed description of HTTPChaos configuration, see Field description below.

Notes

Before injecting the faults supported by HTTPChaos, note the followings:

There is no control manager of Chaos Mesh running on the target Pod.
The rules will affect both of clients and servers in the Pod, if you want to affect only one side, please refer to the specify side section.
HTTPS accesses should be disabled, because injecting HTTPS connections is not supported currently.
For HTTPChaos injection to take effect, the client should avoid reusing TCP socket. This is because HTTPChaos does not affect the HTTP requests that are sent via TCP socket before the fault injection.
Use non-idempotent requests (such as most of the POST requests) with caution in production environments. If such requests are used, the target service may not return to normal status by repeating requests after the fault injection.

Create experiments using Chaos Dashboard

Open Chaos Dashboard, and click NEW EXPERIMENT on the page to create a new experiment:
In the Choose a Target area, choose HTTP FAULT and select a specific behavior, such as RESPONSE ABORT. Then fill out specific configurations.
Submit the experiment.

In the example above, you have configured injecting the "Response abort" fault into all requests of Port 80.

Create experiments using YAML files

Chaos Mesh also supports using YAML configuration files to create HTTPChaos experiments. In a YAML file, you can simulate either one HTTP fault type or a combination of different HTTP fault types.

Example of `abort`

Write the experimental configuration to the http-abort-failure.yaml file as the example below:
```
apiVersion: chaos-mesh.org/v1alpha1
kind: HTTPChaos
metadata:
  name: test-http-chaos
spec:
  mode: all
  selector:
    labelSelectors:
      app: nginx
  target: Request
  port: 80
  method: GET
  path: /api
  abort: true
  duration: 5m
```
Based on this configuration example, Chaos Mesh will inject the abort fault into the specified pod for 5 minutes. During the fault injection, the GET requests sent through port 80 in the /api path of the target Pod will be interrupted.
After the configuration file is prepared, use kubectl to create the experiment:
```
kubectl apply -f ./http-abort-failure.yaml
```

Example of fault combinations

Write the experimental configuration to http-failure.yaml file as the example below:

apiVersion: chaos-mesh.org/v1alpha1
kind: HTTPChaos
metadata:
  name: test-http-chaos
spec:
  mode: all
  selector:
    labelSelectors:
      app: nginx
  target: Request
  port: 80
  method: GET
  path: /api/*
  delay: 10s
  replace:
    path: /api/v2/
    method: DELETE
  patch:
    headers:
      - ['Token', '<one token>']
      - ['Token', '<another token>']
    body:
      type: JSON
      value: '{"foo": "bar"}'
  duration: 5m

Based on this configuration example, Chaos Mesh will inject the delay fault, replace fault, and patch fault consecutively.

After the configuration file is prepared, use kubectl to create the experiment:
```
kubectl apply -f ./http-failure.yaml
```

Field description

Description for common fields

Common fields are meaningful when the target of fault injection is Request or Response.

Parameter	Type	Description	Default value	Required	Example
`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).		yes	`one`
`value`	string	Provides parameters for the `mode` configuration depending on the value of `mode`.		no	1
`target`	string	Specifies whether the target of fault injuection is `Request` or `Response`. The `target`-related fields should be configured at the same time.		yes	Request
`port`	int32	The TCP port that the target service listens on.		yes	80
`path`	string	The URI path of the target request. Supports Matching wildcards.	Takes effect on all paths by default.	no	/api/*
`method`	string	The HTTP method of the target request method.	Takes effect for all methods by default.	no	GET
`request_headers`	map[string]string	Matches request headers to the target service.	Takes effect for all requests by default.	no	Content-Type: application/json
`abort`	bool	Indicates whether to inject the fault that interrupts the connection.	false	no	true
`delay`	string	Specifies the time for a latency fault.	0	no	10s
`replace.headers`	map[string]string	Specifies the key pair used to replace the request headers or response headers.		no	Content-Type: application/xml
`replace.body`	[]byte	Specifies request body or response body to replace the fault (Base64 encoded).		no	eyJmb28iOiAiYmFyIn0K
`patch.headers`	[][]string	Specifies the attached key pair of the request headers or response headers with patch faults.		no	- [Set-Cookie, one cookie]
`patch.body.type`	string	Specifies the type of patch faults of the request body or response body. Currently, it only supports `JSON`.		no	JSON
`patch.body.value`	string	Specifies the fault of the request body or response body with patch faults.		no	`{"foo": "bar"}`
`duration`	string	Specifies the duration of a specific experiment.		yes	30s
`scheduler`	string	Specifies the scheduling rules for the time of a specific experiment.		no	5 * * * *
`tls.secretName`	string	SecretName represents the name of required secret resource. The secrete must combined with data `{"tls.certName":cert, "tls.KeyName":key, "tls.caName":ca}`		no	"http-tls-scr"
`tls.secretNamespace`	string	SecretNamespace represents the namespace of required secret resource,should be the same with deployment/chaos-controller-manager in most cases		no	"chaos-mesh"
`tls.certName`	string	CertName represents the data name of cert file in secret, `tls.crt` for example		no	"tls.crt"
`tls.KeyName`	string	KeyName represents the data name of key file in secret, `tls.key` for example		no	"tls.key"
`tls.caName`	string	CAName represents the data name of ca file in secret, `ca.crt` for example		no	"ca.crt"

:::note

When creating experiments with YAML files, replace.body must be the base64 encoding of the replacement content.
When creating experiments with the Kubernetes API, there is no need to encode the replacement content, just convert it to []byte and put it into the httpchaos.Spec.Replace.Body field. The following is an example:

httpchaos.Spec.Replace.Body = []byte(`{"foo": "bar"}`)

:::

Description for `target`-related fields

`Request`-related fields

The Request field is a meaningful when the target set to Request during the fault injection.

Parameter	Type	Description	Required	Example
`replace.path`	string	Specifies the URI path used to replace content.	no	/api/v2/
`replace.method`	string	Specifies the replaced content of the HTTP request method.	no	DELETE
`replace.queries`	map[string]string	Specifies the replaced key pair of the URI query.	no	foo: bar
`patch.queries`	[][]string	Specifies the attached key pair of the URI query with patch faults.	no	- [foo, bar]

`Respond`-related fields

The Response is a meaningful when the target set to Response during the fault injection.

Parameter	Type	Description	Default value	Required	Example
`code`	int32	Specifies the status code responded by `target`.	Takes effect for all status codes by default.	no	200
`response_headers`	map[string]string	Matches request headers to `target`.	Takes effect for all responses by default.	no	Content-Type: application/json
`replace.code`	int32	Specifies the replaced content of the response status code.		no	404

Specify side

The rules will affect both of clients and servers in the Pod by default, but you can affect only one side by selecting the request headers.

This section provides some examples to specify the affected side, you can adjust the header selector in rules depend on your particular cases.

Client side

To inject faults into clients in the Pod without affecting servers, you can select the request/response by the Host header in the request.

For example, if you want to interrupt all requests to http://example.com/, you can apply the following YAML config:

apiVersion: chaos-mesh.org/v1alpha1
kind: HTTPChaos
metadata:
  name: test-http-client
spec:
  mode: all
  selector:
    labelSelectors:
      app: some-http-client
  target: Request
  port: 80
  path: '*'
  request_headers:
    Host: 'example.com'
  abort: true

Server side

To inject faults into servers in the Pod without affecting clients, you can also select the request/response by the Host header in the request.

For example, if you want to interrupt all requests to your server behind service nginx.nginx.svc, you can apply the following YAML config:

apiVersion: chaos-mesh.org/v1alpha1
kind: HTTPChaos
metadata:
  name: test-http-server
spec:
  mode: all
  selector:
    labelSelectors:
      app: nginx
  target: Request
  port: 80
  path: '*'
  request_headers:
    Host: 'nginx.nginx.svc'
  abort: true

In other cases, especially when injecting the inbound request from outside, you may select the request/response by the X-Forwarded-Host header in the request.

For example, if you want to interrupt all requests to your server behind a public gateway nginx.host.org, you can apply the following YAML config:

apiVersion: chaos-mesh.org/v1alpha1
kind: HTTPChaos
metadata:
  name: test-http-server
spec:
  mode: all
  selector:
    labelSelectors:
      app: nginx
  target: Request
  port: 80
  path: '*'
  request_headers:
    X-Forwarded-Host: 'nginx.host.org'
  abort: true

TLS

To inject faults inside connection base on TLS, user should use TLS mode. Our proxy play a proxy role here, so in TLS people both need to act as a remote server with a trustful CA , but also need to act as a client trust the server with some ca.

So in the secret data blow user need to create its' TLS keys & CA & CRT on their own.

{
  "tls.certName":cert,
  "tls.KeyName":key,
  "tls.caName":ca
}

If user need to create a new TLS server and inject the connection to it, they should:

Create their own root CA's private key and root CA's certificate:

openssl req -newkey rsa:4096  -x509  -sha512  -days 365 -nodes -out ca.crt -keyout ca.key

Create their server's Certificate Signing Request:

openssl genrsa -out server.key 2048
openssl req -new -key server.key -out server.csr

Write an extension file server.ext for the server like:

authorityKeyIdentifier=keyid,issuer
basicConstraints=CA:FALSE
keyUsage = digitalSignature, nonRepudiation, keyEncipherment, dataEncipherment
subjectAltName = @alt_names

[alt_names]
IP.1 = X.X.X.X

Generate certificate of server:

openssl x509 -req -in server.csr -CA ca.crt -CAkey ca.key -CAcreateserial -out server.crt -days 365 -sha256 -extfile server.ext

Add CA ca.crt to client.
Put server.key, server.crt, ca.crt into a secrete and give it to TLS mode.

If user need to inject a client , they should act the proxy of HTTP Chaos like the remote server , you should just edit server.ext above to the specify domain.

Example:

subjectAltName = @alt_names

[alt_names]
DNS.1 = *.domain.com
IP.1 = xxx.xxx.xxx.xxx

Local debugging

If you are not sure of the effects of certain fault injections, you can also test the corresponding features locally using rs-tproxy. Chaos Mesh also provides HTTP Chaos by using chaos-tproxy.

14 KiB Raw Permalink Blame History