JSON API Grafana Datasource

The JSON Datasource executes requests against arbitrary backends and parses JSON response into Grafana dataframes.

Contents

• Installation
• Setup
• API
  • /search
  • /query
  • /annotations
  • /tag-keys
  • /tag-values
• Development Setup

Installation

To install this plugin using the grafana-cli tool:

 grafana-cli plugins install simpod-json-datasource

See here for more information.

Setup

When adding datasource add your API endpoint to the URL field. That's where datasource will make requests to.

API

An OpenAPI definition is defined at openapi.yaml.

To work with this datasource the backend needs to implement 4 endpoints:

• GET / with 200 status code response. Used for "Test connection" on the datasource config page.
• POST /search returning available metrics when invoked.
• POST /query returning metrics based on input.
• POST /annotations returning annotations.

Those two urls are optional:

• POST /tag-keys returning tag keys for ad hoc filters.
• POST /tag-values returning tag values for ad hoc filters.

/search

POST /search

Grafana issues this request on

1. Variables > New/Edit page. Query field value is passed in a body as

{ "target": "query field value" }

2. Panel > Queries page. Format As and Metric values are passed in a body as

{ "type": "timeseries", "target": "upper_50" }

The way you handle those values is up to you.

The response body can either contain an array or a map.

Example array response:

["upper_25","upper_50","upper_75","upper_90","upper_95"]

Example map response:

[ { "text": "upper_25", "value": 1}, { "text": "upper_75", "value": 2} ]

/query

POST /query

Example timeseries request:

{
  "panelId": 1,
  "range": {
    "from": "2016-10-31T06:33:44.866Z",
    "to": "2016-10-31T12:33:44.866Z",
    "raw": {
      "from": "now-6h",
      "to": "now"
    }
  },
  "rangeRaw": {
    "from": "now-6h",
    "to": "now"
  },
  "interval": "30s",
  "intervalMs": 30000,
  "maxDataPoints": 550,
  "targets": [
     { "target": "Packets", "refId": "A", "type": "timeseries", "data": { "additional": "optional json" } },
     { "target": "Errors", "refId": "B", "type": "timeseries" }
  ],
  "adhocFilters": [{
    "key": "City",
    "operator": "=",
    "value": "Berlin"
  }]
}

Example timeseries response (metric value as a float , unixtimestamp in milliseconds):

[
  {
    "target":"pps in",
    "datapoints":[
      [622,1450754160000],
      [365,1450754220000]
    ]
  },
  {
    "target":"pps out",
    "datapoints":[
      [861,1450754160000],
      [767,1450754220000]
    ]
  },
  {
    "target":"errors out",
    "datapoints":[
      [861,1450754160000],
      [767,1450754220000]
    ]
  },
  {
    "target":"errors in",
    "datapoints":[
      [861,1450754160000],
      [767,1450754220000]
    ]
  }
]

The relation between target in request and response is 1:n. You can return multiple targets in response for one requested target.

Example table response to be returned if the metric selected is "type": "table":

[
  {
    "columns":[
      {"text":"Time","type":"time"},
      {"text":"Country","type":"string"},
      {"text":"Number","type":"number"}
    ],
    "rows":[
      [1234567,"SE",123],
      [1234567,"DE",231],
      [1234567,"US",321]
    ],
    "type":"table"
  }
]

Additional data

Sending additional data for each metric is supported via the Additional JSON Data input field that allows you to enter JSON.

For example when { "additional": "optional json" } is entered into Additional JSON Data input, it is attached to the target data under "data" key:

{ "target": "upper_50", "refId": "A", "type": "timeseries", "data": { "additional": "optional json" } }

You can also enter variables:

/annotations

POST /annotations

The JSON request body looks like this:

{
  "range": {
    "from": "2016-04-15T13:44:39.070Z",
    "to": "2016-04-15T14:44:39.070Z"
  },
  "rangeRaw": {
    "from": "now-1h",
    "to": "now"
  },
  "annotation": {
    "name": "deploy",
    "datasource": "JSON Datasource",
    "iconColor": "rgba(255, 96, 96, 1)",
    "enable": true,
    "query": "#deploy"
  },
   "variables": []
}

Grafana expects a response containing an array of annotation objects.

Field explanation:

• text - Text for the annotation. (required)
• title - The title for the annotation tooltip. (optional)
• isRegion - Whether is region. (optional) (http://docs.grafana.org/reference/annotations/#adding-regions-events)
• time - Time since UNIX Epoch in milliseconds. (required)
• timeEnd - Time since UNIX Epoch in milliseconds (required if isRegion is true )
• tags - Tags for the annotation. (optional)

[
  {
    "text": "text shown in body",
    "title": "Annotation Title",
    "isRegion": true,
    "time": "timestamp",
    "timeEnd": "timestamp",
    "tags": ["tag1"]
  }
]

Note: If the datasource is configured to connect directly to the backend, you also need to implement OPTIONS /annotations that responds with the correct CORS headers:

Access-Control-Allow-Headers:accept, content-type
Access-Control-Allow-Methods:POST
Access-Control-Allow-Origin:*

/tag-keys

POST /tag-keys

Example request body

{ }

The tag keys api returns:

[
    {"type":"string","text":"City"},
    {"type":"string","text":"Country"}
]

/tag-values

POST /tag-values

Example request body

{"key": "City"}

The tag values api returns:

[
    {"text": "Eins!"},
    {"text": "Zwei"},
    {"text": "Drei!"}
]

Development Setup

This plugin requires node 6.10.0. Use of Yarn is encouraged to build.

yarn install
yarn run build