InfluxDB 1.2 Released with Subqueries and 50% Better Write Throughput on Larger Hardware

By Paul Dix / Company, Press Releases
Jan 24, 2017

Today we’re excited to announce the release of InfluxDB 1.2 along with the other components of the TICK stack, an open source, end-to-end solution for solving monitoring and analytics problems with time series data. This release of the stack brings some exciting new developments in both our open source and commercial projects. It comes about three months after the release of 1.1, which is exactly what we were aiming for. Read on for more details about improvements in InfluxDB and Kapacitor specifically.

InfluxDB 1.2 Improvements

This release made significant improvements in reducing lock contention in the write path. This means that on larger hardware we more effectively use all available CPU cores, which has resulted in a 50% (or better) gain on total write throughput. For example, we were able to write ~2M values/sec on a C4 36 core Amazon EC2 instance. For the Golang audience the PR might be interesting reading.

With this release, InfluxDB introduces support for subqueries. Specifically, you can now execute queries like the sum of derivatives, which is useful for things like seeing total network bandwidth across a group of machines. This is a first step in terms of our support for subqueries, but we feel it is a very useful step and look forward to community suggestions about broadening our support for subqueries in subsequent releases. You can read more about this new feature in the subquery documentation.

InfluxDB Enterprise, our commercial edition that adds high availability and scale out clustering, now provides  built-in, incremental backup. It’s very useful for customers that have larger clusters or larger databases.

Full details on what’s in the open source release can be found in the InfluxDB changelog for 1.2.

Kapacitor 1.2 Improvements

Kapacitor 1.2 has many additional features and bug fixes. Read on to hear about new ways to window data and an all new alerting API and functionality. See the Kapacitor changelog for full details on the 1.2 release.

In addition to the ability to aggregate data into logical windows based on time, we’ve expanded this further to allow for windows to be defined by counts as well. This provides users with improved flexibility in terms of filtering and organizing large amounts of data.

We’ve added enhanced alert condition handling which now allows users to attach alert handlers to alert topics. The decoupling of the alert condition and alert handling, which was previously performed within TICK Script, eliminates duplication of similar alert handlers and allows for greater flexibility with regards to handling the alerts themselves.

For more information on the new Kapacitor alert topics functionality see the example or look at the Kapacitor alerts API documentation.

Next steps

We’re excited about this release and our continued ability to iterate quickly on the feedback of the open source community. With the four components, Telegraf, InfluxDB, Chronograf, and Kapacitor, we’ve built a complete open source, end-to-end solution for monitoring and working with time series data. We hope you’ll give it a try and we look forward to your feedback.