Mockintosh

You’ve just found a new way of mocking microservices!

An example config that demonstrates the common features of Mockintosh:

services:
  - name: Mock for Service1
    hostname: localhost
    port: 8000
    managementRoot: __admin  # open http://localhost:8001/__admin it in browser to see the UI  
    endpoints:

    - path: "/"  # simplest mock

    - path: "/api/users/{{param}}"  # parameterized URLs
      response: "simple string response with {{param}} included"

    - path: /comprehensive-matching-and-response
      method: POST
      queryString:
        qName1: qValue  # will only match if query string parameter exists
        qName2: "{{regEx '\\d+'}}"  # will require numeric value
      headers:
        x-required-header: someval  # will cause only requests with specific header to work
      body:
        text: "{{regEx '.+'}}"  # will require non-empty POST body
      response:  # the mocked response specification goes below
        status: 202
        body: "It worked"
        headers:
          x-response-header: "{{random.uuid4}}"  # a selection of random/dynamic functions is available
          x-query-string-value: "{{request.queryString.qName2}}"  # request parts can be referenced in response

Mockintosh is a service virtualization tool that’s capable to generate mocks for RESTful APIs and communicate with message queues to either mimic asynchronous tasks or to simulate microservice architectures in a blink of an eye.

The state-of-the-art mocking capabilities of Mockintosh enables software development teams to work independently while building and maintaining a complicated microservice architecture.

Key features:

• Multiple services mocked by a single instance of Mockintosh
• Lenient configuration syntax
• Remote management UI+API
• Request scenarios support with multi-response endpoints and tags
• Mock Actor pattern for Kafka, RabbitMQ, Redis and some other message bus protocols
• GraphQL queries recognizing

In this article we explain how and why Mockintosh was born as a new way of mocking microservices.

Quick Start

Install on Windows

Download the latest installer EXE from releases section and launch it. Follow the steps in wizard to install Mockintosh.

To launch Mockintosh in “quick demo” mode, just start it via Start Menu.

To start Mockintosh with customized configuration and command-line switches, open Command Prompt and start with mockintosh --help command there. See also a section below.

Install on MacOS

Install Mockintosh app on Mac using Homebrew package manager:

$ brew install up9inc/repo/mockintosh

Install on Linux

Install Mockintosh Python package using pip (or pip3 on some machines):

$ pip install -U mockintosh

Use Demo Sample Config

Run following command to generate example.yaml file in the current directory:

$ mockintosh --sample-config example.yaml

then, run that config with Mockintosh:

$ mockintosh example.yaml

And open http://localhost:9999 in your web browser.

You can also issue some CURL requests against it:

curl -v http://localhost:8888/

curl -v http://localhost:8888/api/myURLParamValue123/action

curl -v "http://localhost:8888/someMoreFields?qName1=qValue&qName2=12345" -X POST -H"X-Required-Header: someval" --data "payload"

Download and Use Jinja2-based Sample Config

Once the installation complete, you can start Mockintosh with Jinja2 used as templating language. Use a JSON/YAML configuration as an argument, e.g. example.yaml:

$ mockintosh example.yaml

and you should be seeing a web page like this whenever you visit localhost:8001:

Alternatively, you can run Mockintosh as Docker container:

$ docker run -it -p 8000-8005:8000-8005 -v `pwd`:/tmp up9inc/mockintosh /tmp/config.json

Please note the -p flag used to publish container’s ports and -v to mount directory with config into container.

After server starts, you can issue requests against it. For example, curl -v http://localhost:8000/ would respond hello world. Also, consider opening Management UI in your browser: http://localhost:8000/__admin. Management UI offers visual tools to see available mock endpoints, traffic log and many other features.

Command-line Arguments

The list of command-line arguments can be seen by running mockintosh --help.

If you don’t want to listen all of the services in a configuration file then you can specify a list of service names (name is a string attribute you can set per service):

$ mockintosh example.yaml 'Mock for Service1' 'Mock for Service2'

Using --quiet and --verbose options the logging level can be changed.

Using --bind option the bind address for the mock server can be specified, e.g. mockintosh --bind 0.0.0.0

Using --enable-tags option the tags in the configuration file can be enabled in startup time, e.g. mockintosh --enable-tags first,second

Using --sample-config will cause Mockintosh to write the example configuration file into specified location.

Note: sending SIGHUP to Mockintosh’s process will cause it to re-read configuration file and restart the server.

OpenAPI Specification to Mockintosh Config Conversion (experimental)

Note: This feature is experimental. One-to-one transpilation of OAS documents is not guaranteed.

It could be a good kickstart if you have already an OpenAPI Specification for your API. Mockintosh is able to transpile an OpenAPI Specification to its own config format in two different ways:

CLI Option --convert

Using the --convert one can convert an OpenAPI Specification to Mockintosh config.

JSON output example:

$ wget https://petstore.swagger.io/v2/swagger.json
$ mockintosh swagger.json -c new_config.json json

YAML example:

$ mockintosh swagger.json -c new_config.yaml yaml

Automatic Conversion

If you start Mockintosh with a valid OpenAPI Specification file then it automatically detects that the input is an OpenAPI Specification file:

$ mockintosh swagger.json

and automatically starts itself from that file. Without producing any new files. So you can start to edit this file through the management UI without even restarting Mockintosh.

Interceptors

One can also specify a list of interceptors to be called in <package>.<module>.<function> format using the --interceptor option. The interceptor function get a mockintosh.Request and a mockintosh.Response instance. Here is an example interceptor that for every requests to a path starts with /admin, sets the reponse status code to 403:

import re
from mockintosh import Request, Response


def forbid_admin(req: Request, resp: Response):
    if re.search(r'^\/admin.*$', req.path):
        resp.status = 403

and you would specify such interceptor with a command like below:

$ mockintosh some_config.json --interceptor=mypackage.mymodule.forbid_admin

Instead of specifying a package name, you can alternatively set the PYTHONPATH environment variable to a directory that contains your interceptor modules like this:

$ PYTHONPATH=/some/dir mockintosh some_config.json --interceptor=mymodule.forbid_admin

Note: With interceptors, you can even omit endpoints section from the service config. You will still get all requests to the service into your interceptor.

Request Object

The Request object is exactly the same object defined in here with a minor difference: Instead of accesing the dictonary elements using .<key>, you access them using ['<key>'] e.g. request.queryString['a'].

Response Object

The Response object consists of three fields:

• resp.status holds the HTTP status codes e.g. 200, 201, 302, 404, etc.
• resp.headers is a Python dictionary that you access and/or modify the response headers. e.g. resp.headers['Cache-Control'] == 'no-cache'

• resp.body is usually either str or bytes, but can be anything that is supported

  by tornado.web.RequestHandler.write

  str, bytes or dict e.g. resp.body = 'hello world' or resp.body = {'hello': 'world'}