SplitJoin Microservice

This image receives a single REST request and splits it out into multiple parallel calls. It then gathers the results and returns them as a single response.

This service takes a list of payloads and endpoints and sends the requests to all the endpoints in parallel. It then collects the results and returns the results as a list of responses.

To call splitjoin POST to any URL on the splitjoin server. The requirement to POST specifically to /splitjoin has been removed.

If a GET is done to any URL on the splitjoin server it will return this help.

This image is built on top of the Python:3-slim image. A github project contains the build instructions if a customized image is required.

Authentication

Authentication has the following options:

• No Authentication
  If no Authorization header is passed to the service and username/password are not specified in the query then the call is made without authentication
• Query Specific Basic Authentication
  If username and password are specified in the query then the call is made using basic authentication
• Pass Through Authentication
  If no username/password are specified in the query and the request has an Authorization header then call is made using passed in authorization header. This allow the authentication to be passed through to the target API unchanged.

Base Request Structure

Base request to service has following structure:

    {
        "timeout" : 30,
        "queries" : [
            {
                "id" : "id1",
                "url" : "https://server/path",
                "method" : "POST",
                "username" : "OptionalUsername",
                "password" : "OptionalPassword",
                "headers" : {
                    "header1" : "value1",
                    "header2" : "value2"
                },
                "payload" : {
                    "any" : "payload",
                    "goes" : "here"
                }
            },
            {
                "id" : "id2",
                "url" : "https://server/path",
                "payload" : {
                    "any" : "payload",
                    "goes" : "here"
                }
            },
            {
                "id" : "id3",
                "url" : "https://server/path",
                "method" : "GET",
                "params" : {
                    "param1" : "value1",
                    "param2" : "value2"
                }
            }
        ]
    }

JSon Request Object

• timeout
  Integer: Timeout in seconds for each individual request.
• queries
  Array: Individual calls to execute.
  • id
    String: Unique identifier used by client to correlate the query with the response.
  • url
    URL: Target API to invoke, currently it will always be invoked with a POST.
  • method
    Optional String: HTTP method.
  • username
    Optional String: Userename for basic authentication to API.
  • password
    Optional String: Password for basic authentication to API.
  • headers
    Optional Dictionary: Collection of HTTP headers
    • header
      String: Header name
    • value
      String: Header value
  • payload
    JSon: Payload to be sent to the target API. It can be a primitive or a compound json type.
  • params
    Optional Dictionary: Collection of HTTP Parameters
    • param
      String: Param name
    • value
      String: Param value

Response Structure

Response from service has following structure:

    [
        {
            "start": "2020-05-12 02:23:03.839570",
            "id": "id1",
            "response": {
                "any" : "response",
                "goes" : "here"
            },
            "headers": {
                "header1" : "value1",
                "header2" : "value2"
            },
            "status": 200,
            "end": "2020-05-12 02:23:27.597740"
        },
        {
            "start": "2020-05-12 02:23:03.843628",
            "id": "id2",
            "response": {
                "any" : "response",
                "goes" : "here"
            },
            "headers": {
                "header1" : "value1",
                "header2" : "value2"
            },
            "status": 200,
            "end": "2020-05-12 02:23:30.594746"
        }
    ]

JSon Response Object

• Array: Individual responses.
  • start
    String: Date time call was made
  • id
    String: Unique identifier used by client to correlate the response with the query.
  • response
    JSon: Response received from the target API. It can be a primitive or a compound json type.
  • headers
    Dictionary: Received HTTP Headers
    • header
      String: Header name
    • value
      String: Header value
  • status
    Integer: HTTP response code received from call
  • end
    String: Date time response was received

Deployment

There are a number of ways to deploy the service:

• Run standalone using "python3 splitjoin.py". Note that this requires aiohttp and markdown packages.
• Run as a docker image using "docker run -p 8080:8080 docker.io/antonyjreynolds/splitjoin:latest"
• Deploy to Kubernetes using deployment.yaml, if using Oracle OCI then the service deployment service-oci-lb.yaml will create a load balancer.

Using with OIC

A sample package Recalls.par is provided that shows how to use splitjoin from OIC. When setting up a connection to SplitJoin it is a good idea to set the security credentials to be OIC credentials. This can be either Basic Auth or OAuth. This will allow the SplitJoin to call OIC integrations without providing any additional credentials in the invoke.