Section IV - database

CouchDB API

rust-daas.postman_collection.json

history.json

When following a true Microservice architecture, it is important to remember that each microsservice has its own data source. This ensures higher availablity of the service and supports agility of the architecture by reduces inter-dependencies. All too often, the microservice's application layer are constructed as independently executables, while their data source(s) remain tightly coupled.

Since the Order Status History RESTful service only the consumer of the reporting data, (and source of record satellite ) the database will be synchronized in realtime uni-directionally from the provisioning datbasase (source of record).

Creating the Database

What's nice about CouchDB is that the platform doesn't require the installation of client software or driver. All database activities can either be done (1) manually through the Fauxton Console (http://couchdb-servcer:5984/_utils/) or (2) programmatically through CouchDB's RESTful API (http://docs.couchdb.org/en/latest/api/index.html).

Let's create a new databasse named consuming using the RESTful API. This database will only house a copy of the reporting data that's required for the Order Status History RESTful service.

HTTP Request

curl -X PUT \
  'http://localhost:5984/consuming?q=8&n=3&partitioned=false&partitioned%20=false' \
  -H 'Accept: */*' \
  -H 'Content-type: application/json' \
  -H 'cache-control: no-cache'

HTTP Response

{
    "ok": true
}

Replicating the Realtime Data

Now that we have a database for the Order Status History RESTful service, we will setup the uni-directional realtime sychronization.

HTTP Request

curl -X POST \
  http://localhost:5984/_replicate \
  -H 'Accept: application/json' \
  -H 'Accept-Encoding: gzip, deflate' \
  -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ=' \
  -H 'Cache-Control: no-cache' \
  -H 'Connection: keep-alive' \
  -H 'Content-Length: 88' \
  -H 'Content-Type: application/json' \
  -H 'Host: localhost:5984' \
  -H 'cache-control: no-cache' \
  -d '{
    "continuous" : true,
    "source": "provisioning",
    "target": "consuming"
}'

HTTP Response

{
    "ok": true,
    "_local_id": "bedfef67af6c3e82a425fece02aee9fc+continuous"
}

Just verify that the documents are now replicated to the consuming database, make a call to the CouchDB API to get a list of all the documents in the consuming databaase.

HTTP Request

curl -X GET \
  http://localhost:5984/consuming/_all_docs \
  -H 'Accept: application/json' \
  -H 'Accept-Encoding: gzip, deflate' \
  -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ=' \
  -H 'Cache-Control: no-cache' \
  -H 'Connection: keep-alive' \
  -H 'Host: localhost:5984' \
  -H 'cache-control: no-cache'

HTTP Response

{
    "total_rows": 2,
    "offset": 0,
    "rows": [
        {
            "id": "history|status|iStore|8003",
            "key": "history|status|iStore|8003",
            "value": {
                "rev": "3-9b8f57b145ddfea0e3fec0147a6cb835"
            }
        },
        {
            "id": "history|status|iStore|8004",
            "key": "history|status|iStore|8004",
            "value": {
                "rev": "1-1f4e25879ee846bc6ca4d5a00bbfb472"
            }
        }
    ]
}

Creating a Data View

Now that we have data in the consuming databaase, we should build a data interface (View) for the Order Status History RESTful service to interact with to retreive the data. We will do this by creating a _design document using the CouchDB API

HTTP Request

curl -X PUT \
  http://localhost:5984/consuming/_design/history \
  -H 'Accept: application/json' \
  -H 'Accept-Encoding: gzip, deflate' \
  -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ=' \
  -H 'Cache-Control: no-cache' \
  -H 'Connection: keep-alive' \
  -H 'Content-Length: 504' \
  -H 'Content-Type: application/json' \
  -H 'Host: localhost:5984' \
  -H 'cache-control: no-cache' \
  -d '{
    "_id": "_design/history",
    "views": {
      "status-duration": {
        "map": "function (doc) {\n  \n  for(s = 0; s < doc.data_obj.order_status.length; s++) {\n    var duration = doc.data_obj.order_status[s].timestamp - doc.data_obj.timestamp;\n    emit(doc.data_obj.order_status[s].name, duration/60);\n  }\n}",
        "reduce": "function (keys, values, rereduce) {\n    avg = Math.round(sum(values)/values.length);\n    return avg;\n}"
      }
    },
    "language": "javascript"
}'

HTTP Response

{
    "ok": true,
    "id": "_design/history",
    "rev": "1-61a4a24d9e3a86cf154665a0588211ea"
}