Skip to content

Latest commit

 

History

History
298 lines (188 loc) · 13.7 KB

README.md

File metadata and controls

298 lines (188 loc) · 13.7 KB

Pipeline (Github Action)

CI If PUSH or PR to master the following steps will run:

  • yarn install
  • yarn build
  • yarn test
  • Create github pre-release

Worldmap Panel Plugin for Grafana

The Worldmap Panel is a tile map of the world that can be overlaid with circles representing data points from a query. It can be used with time series metrics, with geohash data from Elasticsearch or data in the Table format.

Worldmap

How The Worldmap Works (Theory and Examples)

The Worldmap panel needs two sources of data:

  • a location (latitude and longitude)
  • data that has a link to a location

The data comes from a database query: Prometheus, InfluxDB, Graphite, Elasticsearch, MySQL etc. It can be in the Time Series format or in the Table format.

Time Series Format

If it is in the Time Series format then the metric name needs to match a key from a list of locations. That key is usually a country code or city name. The list of locations can come from a file or an HTTP endpoint.

The list of locations can be provided in several ways:

  • json files with locations and their coordinates (the plugin includes list for countries and US states)
  • json endpoints that return a list of locations and their coordinates

Time Series data contains a timestamp, a metric name and a numeric value. In other words, a typical query for a time series database. Here is some time series data from Graphite:

[
  {"target": "SE", "datapoints": [[183255.0, 1529755200]]},
  {"target": "US", "datapoints": [[192224.0, 1529755200]]}
]

Location data should be in the JSON format and should be an array of JSON objects with four properties:

[
  {
    "key": "SE",
    "latitude": 60.128161,
    "longitude": 18.643501,
    "name": "Sweden"
  },
  {
    "key": "US",
    "latitude": 37.09024,
    "longitude": -95.712891,
    "name": "United States"
  }
]

The Worldmap will then match the metric name (target in the example data) with a key field from the location data. With this example data there will be two circles drawn on the map, one for Sweden and one for the United States with values 183255 and 192224.

Table Format

If the data is in the Table format then it should have a column that is a geohash or two columns that contain the latitude and longitude (together with the columns for the data).

Table data is tabular data with columns and rows. Here is an example of Table data from InfluxDB:

"series": [
  {
    "name": "logins.count",
    "tags": {
      "geohash": "9wvfgzurfzb"
    },
    "columns": [
      "time",
      "metric"
    ],
    "values": [
      [
        1529762933815,
        75.654324173059
      ]
    ]
  }
]

This query contains both data (the value 75.654324173059) and a location (the geohash 9wvfgzurfzb which is in Colorado). So using these, the Worldmap panel will draw one circle in Colorado, USA with the value 75.654324173059.

Time Series Data as the Data Source

Supported Databases:

  • Graphite
  • InfluxDB
  • OpenTSDB
  • Prometheus
  • MySQL
  • Postgres
  • MSSQL
  • Elasticsearch

The following location files are included in the plugin:

  • Countries (2 letter codes)
  • Countries (3 letter codes)
  • US states

Alternatively, you can provide your own location lists by using:

  • A JSON endpoint that returns a list of locations
  • A JSONP endpoint that returns a list of locations

This works by matching country codes (like US or GB or FR) or US state codes (TX or NY) to a metric name. If a metric name matches a country in the list of countries then a circle will be drawn at that location.

If you want to match to other data than countries or states, then you will have to provide custom location data. The current way to do that is via a JSON endpoint that returns a json file with location data (See Map Data Options)

The size of the circle depends on the value of the matched metric. Circle size is relative e.g. if you have 3 countries with values 1, 2 and 3 or 100, 200 and 300 then you will get one small circle, one medium circle and one large circle.

Time Series - Graphite and InfluxDB

Here are some examples of Time Series Queries

Graphite Query

Use the aliasByNode function to point to the field containing the country code. See the image below for an example of a graphite query.

Graphite Query for Countries

Example dashboard for Worldmap with Graphite queries on the Grafana play site.

InfluxDB Query

The Group By clause should be the country code and an alias is needed too. The alias should be in the form $tag_<field name>.

Influx Query for Countries

Elasticsearch Query for Countries

Use a Group By clause on the field containing the country code and a Then by clause with Date Histogram by @timestamp (or corresponding date field).

Elasticsearch Query for Countries

Map Data Options for Time Series Data

Under the Worldmap tab, choose either the countries or states option.

Worldmap Options for Countries

Using a JSON endpoint to return a custom list of locations:

Worldmap Options for JSON

The endpoint used here is for the demo version of worldPing - https://worldpingdemo.grafana.net/api/plugin-proxy/raintank-worldping-app/api/v2/probes/locations. If you have your own endpoint defined it must be reachable from the client side, as it is approached by client's browser.

Using a JSONP endpoint (if you need to wrap the JSON to get around CORS problems):

Worldmap Options for JSONP

For some details on troubleshooting JSON/JSONP check #47.

Geohashes as the Data Source

Supported Databases:

  • ElasticSearch

The Geo-point data type with geohash indexing in Elasticsearch can also be used as a datasource for the worldmap panel. Grafana has a new bucket aggregate for Elasticsearch queries - Geo Hash Grid that allows grouping of coordinates. The Geo Hash Grid has a precision option where 1 is the highest level and 7 is the lowest.

Elasticsearch Query for Worldmap

Three fields need to be provided by the ElasticSearch query:

  • A metric. This is free text and should match the aggregation used (Count, Average, Sum, Unique Count etc.)
  • Location Name (optional - geohash value will be shown if not chosen)
  • geo_point field that provides the geohash value.

Elasticsearch Query for Worldmap

Table Data as the Data Source

Supported Databases:

  • InfluxDB
  • Elasticsearch
  • MySQL, Postgres, MSSQL
  • Any database that can return data in Table Format

If a datasource can return Table Data then on the Metrics tab in Grafana choose the FORMAT AS Table option.

Table Data with a Geohash Column

Similar to the Elasticsearch query above, 3 fields are expected (2 of them are mandatory)

  • A metric field. This is used to give the circle a value and determines how large the circle is.
  • A geohash field. This is used to calculate where the circle should be drawn.
  • an optional location name field (shown in the mouse over). Used to label each circle on the map. If it is empty then the geohash value is used as the label.

The field mappings have to be specified on the Worldmap settings tab.

Example influxdb query

Table Data with Latitude and Longitude Columns

The Table Data format also works with two columns for latitude and longitude instead of a geohash column.

  • A metric field. This is used to give the circle a value and determines how large the circle is.
  • Latitude/Longitude Fields. These are used to calculate where the circle should be drawn.
  • an optional location name field (shown in the mouse over). Used to label each circle on the map. If it is empty then the geohash value is used as the label.

JSON result as the Data Source

Supported Databases:

It supports any datasource capable of generating a JSON response with a a custom list of locations (the same format that for the JSON enpoint).

Map Data Options

Location Data

There are four ways to provide data for the worldmap panel:

  • countries: This is a list of all the countries in the world. It works by matching a country code (US, FR, AU) to a node alias in a time series query.
  • states: Similar to countries but for the states in USA e.g. CA for California
  • geohash: An ElasticSearch query that returns geohashes.
  • json: A json endpoint that returns custom json. Examples of the format are the countries data used in first option or this list of cities.
  • jsonp: A jsonp endpoint that returns custom json wrapped as jsonp. Use this if you are having problems with CORS.
  • table: This expects the metric query to return data points with a field named geohash or two fields/columns named latitude and longitude. This field should contain a string in the geohash form. For example: London -> "gcpvh3zgu992".

Aggregation

If you chose countries or table as the source of the location data then you can choose an aggregation here: avg, total etc.

For Graphite, be aware that the default value for Max Data Points is 1. This is to aggregate data points per country to one value in the most accurate way. This will by default, consolidate by average. To change the consolidation, use the consolidateBy function like so:

Graphite Consolidate By

Or just remove the 1 from the Max Data Point field and use the consolidation functions in Map Data Options (though depending on the timerange and amount of data points, this will be usually less accurate due to Graphite consolidation).

Graphite Max Data Points

ES Metric/Location Name/geo_point Field

Three fields need to be provided by the ElasticSearch query. They are text fields and should be the field names from the query under the Metrics tab.

  • The Metric in Elasticsearch is one of Count, Average, Sum etc.
  • Location Name is the field that gives the circle a name. If it is blank, then the geohash value is shown in the popover instead of the location.
  • geo_point is the GeoHashGrid field that provides the geohash value.

Map Visual Option Settings

Center

This settings configures the default center of the map. There are 5 centers to choose from or you can choose a custom center or last GeoHash center..For a custom center there are two fields: latitude and longitude. Examples of values are 37.09024, -95.712891 for the center of the US or 55.378051, -3.435973 for Great Britain. Last GeoHash center will centered the map on the last GeoHash received from the data.

Initial Zoom

The initial zoom factor for the map. This is a value between 1 and 18 where 1 is the most zoomed out.

Min Circle Size

This is minimum size for a circle in pixels.

Max Circle Size

This is the maximum size for a circle in pixels. Depending on the zoom level you might want a larger or smaller max circle size to avoid overlapping.

Unit

The Unit is shown in the popover when you hover over a circle. There are two fields the singular form and the plural form. E.g. visit/visits or error/errors

Show Legend

Shows/hide the legend on the bottom left that shows the threshold ranges and their associated colors.

Threshold Options

Thresholds control the color of the circles.

If one value is specified then two colors are used. For example, if the threshold is set to 10 then values under 10 get the first color and values that are 10 or more get the second color.

The threshold field also accepts 2 or more comma-separated values. For example, if you have 2 values that represents 3 ranges that correspond to the three colors. For example: if the thresholds are 70, 90 then the first color represents < 70, the second color represents between 70 and 90 and the third color represents > 90.

Makefile

Added a Makefile. Available commands are:

make build

This command will cleanup the dist folder, the dependencies and run the npm build command.

CHANGELOG

The latest changes can be found here: CHANGELOG.md