Data / Development / Sponsored / Contributed

Getting Started with JavaScript and InfluxDB

20 Aug 2021 7:27am, by

Telegraf is the preferred way to collect data for InfluxDB. Though in some use cases, client libraries are better, such as when parsing a stream of server-side events. In this tutorial, you’ll learn how to read a data stream, store it as a time series into InfluxDB and run queries over the data using InfluxDB’s JavaScript client library.

All the code in this tutorial is available for free in this repo on GitHub.

What Is a Time-Series Database?

Nicolas Bohorquez
Nicolas is a data architect at Merqueo. He has been part of development teams in a handful of startups and has founded three companies in the Americas. He is passionate about the modeling of complexity and the use of data science to improve the world.

A time-series database is a specialized type of data store focused on providing tools to store and query data that has a dimension measured as a time unit. Good examples include the temperature at a particular minute of every hour, the price of a stock in the stock market or the number of cars in one area during rush hour.

There are many examples of time-based datasets. Not all types of data are suitable for time-series databases. One example is the Iris dataset used for training classification problems in the machine learning space. Nor is the Titanic dataset, which is used in forecasting.

Time-series databases are often used to analyze application logs and collect sensor data. Applications and sensors produce streams of data constantly, with attributes different from the time-based dimension. In the case of this tutorial, the data source is the event stream of the recent changes provided by the Wikimedia Foundation. This data follows the Server-Sent Event and can be consumed directly via HTTP.

The sample Node.js CLI application consists of two components. The first consumes stream messages and writes them as data points in the InfluxDB database. Second, there’s a reader that queries the database to render the results of the query as a simple line chart. Time is rendered in the x-axis, and the value of the series in the y-axis.

Tutorial prerequisites

This example is tested using Ubuntu 20.04 and Node.js v14.17.3 (npm v6.14.13) — installed using the Node Version Manager (NVM). With the many versions of Node.js available, NVM helps to manage and test the code easily. This example also writes and reads data from a local InfluxDB 2.0 database.

If you don’t have a local installation, you can follow the installation guide, then create a sample organization, bucket and token.

Once you have the token created, set the values of the following environment variables with the values of your local installation:

  • INFLUX_ORG
  • INFLUX_BUCKET
  • INFLUX_TOKEN
  • INFLUX_URL

Those variables are read in the env.js file, with some default values.

Installing the Library

The InfluxDB JavaScript client is a standard Node.js module that you can install from the command line:

It can also be used as a dependency on browser with a line of code:

For this example, some other libraries are used for rendering purposes, besides the InfluxDB JavaScript client:

  • asciichart: Console ASCII line charts in pure JavaScript
  • blessed: A curses-like library with a high-level terminal interface API (“curses” is a terminal control library for Unix-like systems)
  • chalk: Terminal string styling done right
  • eventsource: A pure JavaScript implementation of an EventSource client

None of these are required to use the InfluxDB JavaScript client.

Making a Connection

Once dependencies are installed and environment variables are set, you can make a connection to the bucket. For this example, the writer.js and reader.js instantiates an InfluxDB object with the url and token read from the environment:

This object provides methods to instantiate the different API clients, such as Writer or Query.

Inserting Data

To insert data in InfluxDB, you have to follow the line protocol. This defines a structure for the four elements that constitute a data point:

  • Measurement: This is the name of the table in which you are going to insert the data
  • Tag set: A comma-separated set of key = value of data properties
  • Field set: A comma-separated set of key = value of data dimensions, the data can be float (default), integer, uinterger, string, Boolean
  • Time stamp (optional): Unix time stamp

Luckily, the InfluxDB JavaScript client provides a wrapper class to be used with the Writer API to insert one or many data points. The following code shows how to connect to the event source and for each message instantiate a point to be added to the bucket:

Notice that in lines 9–11 the writeApi is instantiated with the organization and bucket environment variables, but also with a precision value. In this case it’s seconds, which defines the granularity of the time stamp dimension of the data to be written; check the other possible values here.

Also, you can set some default tags for each data point to be written — the source of the data in this case. Also notice that the client gives you the option to extend this default tag set with other values; for each data point the user and isBot tags are added.

The field written is the length of the change made (line 23), the writePoint method accepts the Point instance, and then the writeApi instance is flushed. You should always close the writeApi in order to flush pending changes and close pending scheduled retry executions. The writer.js runs indefinitely for each message pushed by the event source.

Querying Data

Once the data is in the bucket, the InfluxDB JavaScript client provides another API client to query the data. In this example, the InfluxDB object runs a query that returns the number of data points grouped by the isBot tag in the last 10 seconds:

The query is written in the Flux functional data scripting language. This is designed for querying, analyzing and acting on data over InfluxDB 2.0. The previous language, the InfluxDB SQL-like query language, is still supported at the /query compatibility endpoint of the API, but the recommendation is to use the full power of the new language.

As you can see, the query defines the bucket, range of data, filters to be applied, grouping columns and data functions to be applied. The following function shows you how to run the query against the QueryApi:

For each row returned, the data is parsed into an object and passed to a function that stores it in an array based on the isBot attribute. The two arrays of the series, bots and humans, are then rendered using the asciichart library. The full code for the reader is shown below:

Additional documentation and functionality

The InfluxDB JavaScript client also provides additional functionality (like the HealthAPI, among other wrapper classes) that simplifies building a production-ready pipeline. Check the client and InfluxDB API documentation for more details.

Conclusion

Time-series databases like InfluxDB provide specialized functionality and tools to store and analyze data points with a high rate of ingestion, such as Internet of Things (IoT) data. This example functions as a useful lens to understand the functionality of the InfluxDB JavaScript client to collect high-volume streaming data. The client is a simple, standard Node.js module that lets you write and read data into or from an InfluxDB instance frictionlessly.

The New Stack is a wholly owned subsidiary of Insight Partners. TNS owner Insight Partners is an investor in the following companies: MADE.

Photo by Christina Morillo from Pexels.

A newsletter digest of the week’s most important stories & analyses.