mockgo-server

mockgo-server is a lightweight http server which can be used to mock http endpoints. mockgo-server is designed for horizontal scaling and feels at home in cloud environments like kubernetes and knative. The main features are:

• Simplicity : easy configuration with human readable yaml files with reasonable defaults
• Scalability : mockgo-server is designed for horizontal scaling. Therefore it can be used in environments with high http traffic ( e.g. for performance/load tests )
• Flexibility : sometimes requirements for mocking requests goes beyond static responses. With templating mechanism and state management it is possible to implement particular logic for building responses.

TL;DR

See Examples.md.

variants

mockgo-server is built with different variants which have their own use cases. The following table gives an overview:

variant	can scale	persistence	use this when you ...
mockgo-standalone	no	no	want a simple setup and you don't need to scale, usually this variant is the starting point
mockgo-grpc	yes	no	have to deal with high incoming traffic, but you don't need durability for states and results
mockgo-redis	yes	yes	have to deal with high incoming traffic and you want a durable storage for states and results or if you want to use mockgo as serverless deployment in knative

install on kubernetes

Kubernetes is the first class environment for mockgo-server. So, if you have kubernetes cluster available, you can start right away with helm charts:

helm repo add mockgo-server https://alitari.github.io/mockgo-server/
helm upgrade mymock mockgo-server/mockgo-server --install

see here for further helm configuration options.

install on knative

If you have a knative cluster available, you can start right away with kn

cat <<EOF > kn-mock.yaml
endpoints:
- request:
    path: "/hello-kn"
  response:
    statusCode: 200
    body: "Hello from knative"
EOF
kubectl create configmap mockgo-config --from-file=kn-mock.yaml
kn service create mockgo --image=alitari/mockgo-standalone:v1.3.0 --mount /mockdir=cm:mockgo-config --env MOCK_DIR=/mockdir --env MOCK_PORT=8080
MOCKGO_URL=$(kn service describe mockgo -o url)
curl -v $MOCKGO_URL/hello-kn

See here for a example with redis.

install on local machine

Define the environment :

MOCKGO_RELEASE_VERSION="v1.3.0"
MOCKGO_VARIANT="standalone" # or "grpc"
MOCKGO_OS="linux" # or "windows"
MOCKGO_ARCH="amd64" # or "arm64"
MOCKGO_NAME=mockgo-${MOCKGO_VARIANT}-${MOCKGO_OS}-${MOCKGO_ARCH}

option 1: download binary excutable

wget https://github.com/alitari/mockgo-server/releases/download/$MOCKGO_RELEASE_VERSION/$MOCKGO_NAME
chmod +x $MOCKGO_NAME
# MOCK_DIR is the path for looking for mockfiles, per default this are files with names matching "*-mock.*"
MOCK_DIR=$(pwd)/test/mocks ./$MOCKGO_NAME

option 2: go install

go install github.com/alitari/mockgo-server/mockgo-$MOCKGO_VARIANT/starter/mockgo@latest
MOCK_DIR=$(pwd)/test/mocks mockgo

option 3: run with docker

MOCK_DIR=$(pwd)/test/mocks
docker run -it -v $MOCK_DIR:/mocks -e MOCK_DIR=/mocks alitari/mockgo-$MOCKGO_VARIANT:$MOCKGO_RELEASE_VERSION 

mockfiles and endpoints

A mockgo-server configuration consist of one or multiple files in yaml format, the so called mockfiles. Each mockfile contains one or multiple endpoints. The endpoint defines criteria which qualify the endpoint to serve an incoming request. This process is called matching. The second configuration part of an endpoint is the definition of the http response.

endpoints:
  - id: "id" # [OPTIONAL] unique string to identify endpoint
    prio: 1  # [OPTIONAL] integer to define precedence of endpoints if more than one endpoint matches
    request: # request defines the matching
      host: "alexkrieg.com" # [OPTIONAL], match to http host
      method: "POST" # [OPTIONAL], match to http method, default is "GET"
      path: "/mypath" # [MANDATORY], match to http request path
      query: # [OPTIONAL] for matching, every key value pair must be part of the http query parameters of the incoming request
        firstQueryParam: value1
        secondQueryParam: value2
      headers: # [OPTIONAL] for matching, every key value pair must be part of the http header values of the incoming request
        Content-Type: "application/json"
        Myheader: myheaderValue
      body: "^{.*}$" # [OPTIONAL] regular expression which match to the request body
    response: # defines the response
      statusCode: 204 # [OPTIONAL], http response code ( see RFC 7231), defaults to "200"
      body: "hello" # [OPTIONAL], response body as string, templates can be used
      bodyFilename: "response.json" # [OPTIONAL], refers to a file which contains the response body, templates can be used in the file
      headers: | # [OPTIONAL],multiline string in form of key: value, templates can be used
        Content-Type: "application/text"

path matching

The form of a http request can be described as a sequence of pathsegments which are separated through a /. In order to achieve flexibility matching the path of an http request, there are 3 special symbols which can be used for the pathsegment defining the request.path:

1. * every pathsegment on this position matches
2. {pathvariable} every pathsegment on this position matches, the value of the pathsegment is available as template variable with name pathvariable
3. ** one or more arbitrary pathsegments match on this position

examples

request.path	incoming request path	match
/foo/*	/bar	➖
/foo/*	/foo	➖
/foo/*	/foo/bar	✅
/foo/*	/foo/bar/foo	➖
/foo/*/bar	/foo/bar	➖
/foo/*/bar	/foo/bar/bar	✅
/foo/{foo}/bar	/foo/bar/bar	✅ {{ RequestPathParams.foo }} resolves to bar
/foo/*/{bar}	/foo/bar/bar/foo	➖
/foo/*/{bar}	/foo/bar/bar	✅ {{ RequestPathParams.bar }} resolves to bar
/foo/**	/foo/	➖
/foo/**	/foo/bar	✅
/foo/**	/foo/bar/1/2/3	✅
/foo/**/{bar}	/foo/bar/1	✅ matches but {{ RequestPathParams.bar }} does not resolve
/foo/**/foo/{bar}	/foo/bar/foo	➖
/foo/**/foo/{bar}	/foo/bar/1/2/foo/3	✅ {{ RequestPathParams.bar }} resolves to 3

creating dynamic responses with go templates

In order to implement program logic, you can use go templates for creating dynamic responses. The templating mechanism is available for the response.body and response.headers attributes and provides following additional functions for:

• common template functionality through the sprig library
• access to the incoming request through the Request variables
• using the key-value store

access attributes of the incoming request

template variable	type
RequestPathParams	map[string]string
RequestUrl	string
RequestPath	string
RequestHost	string
RequestBody	string
RequestBodyJSONData	map[string]interface{}

access the key-value store

There are 2 parameters which identifies a value in the key-value store:

• store is a string which identifies the store
• key is a string which identifies the value in the store

template function name	signature	description
kvStoreGet	func(store, key string) interface{}	get the value from store store, which is stored under the key key
kvStorePut	func(store, key string, value interface{})	store value value under the key key in the store store
kvStoreRemove	func(store, key string)	remove the value under the key key from store store

mockgo-server api

The mockgo-server holds multiple kinds of state:

• current configuration (mock endpoints, logging , etc.)
• storage of incoming requests
• key-value store

All states can be accessed through an REST api. The api is secured with basic auth which can be configured with the environment variables API_USERNAME and API_PASSWORD. In order to avoid conflicts with the mock endpoints, the api is exposed under the path prefix /__. This can be changed with the environment variable API_PATH_PREFIX.

configuration api

method	path	description
POST	/__/reload	reload the mock files from the mock dir

matching api

The request storage has a limited capacity which can be configured with MATCHES_CAPACITY.

method	path	description
GET	/__/matches/{endpointId}	returns all requests which matched to an endpoint
GET	/__/matchesCount/{endpointId}	returns the count of requests which matched to an endpoint, is not limited through capacity
GET	/__/mismatches	returns all requests which didn't match to an endpoint
GET	/__/mismatchesCount	returns the count of all requests which didn't match to an endpoint, is not limited through capacity
DELETE	/__/matches/{endpointId}	deletes storage of all requests which matched to an endpoint
DELETE	/__/mismatches	deletes storage of all requests which didn't match to an endpoint

key-value store api

Using the path /__/kvstore/{store}/{key} you can access the key-value store with the following methods:

• PUT stores the content of the request body under the key key in the store store
• GET returns the content which is stored under the key key in the store store
• DELETE removes the content which is stored under the key key in the store store

prometheus metrics

The mockgo-server exposes besides of the standard go metrics the following application specific counters:

• matches{"endpoint":"<endpointId>"}: Number of matches of an endpoint
• mismatches: Number of requests which did not match to on endpoint

using config reload feature

For local development it is useful to have a way to reload the mock files without restarting the mockgo-server. This can be achieved by sending a POST request to the reload endpoint. This functionality can be combined with a file watcher to automatically reload the mock files when they change. The script scripts/watchmocks.sh implements this functionality.

# watch for changes in test/main and reload the mock files
./scripts/watchmocks.sh test/main

contribute

Learn how to contribute here