Customer Success Story: Smart Flows
Smart Flows was looking for an efficient, well-documented, and easy-to-deploy time series database, which they found in InfluxDB. The company uses InfluxDB to receive a stream of radio emissions that come from Bluetooth and wi-fi chips. InfluxDB then helps Smart Flows to monitor the stream and tell if every one of their customer streams is alive or not. Alerts are plugged on Kapacitor and serve information to their forecasting stack; this transforms the number of radio emissions received to the actual number of people. For this task, Smart Flows is transforming the data stored in InfluxDB via a Continuous Query before serving it to a machine learning (ML) tool built in-house. Smart Flows also uses InfluxDB to monitor internal services.
The company found that InfluxDB has a high ingest, integrates well with other tools, and is performant. They ingest millions of events on an hourly basis, so having a way to easily ingest this stream of information and quickly query its content is key. Defining the retention period easily is crucial for Smart Flows, and InfluxDB performs spectacularly. Although Smart Flows may need to switch to InfluxDB Enterprise at some point in the future, the performance of InfluxDB OSS is astonishing, and the OSS version is enough for their current needs.
Smart Flows CTO, Olivier Hervieu, appreciated that InfluxDB is easy to set up and comes with a comprehensive set of tools (Telegraf, Kapacitor) that allows one to quickly set up a comprehensive monitoring suite. The query language helps to onboard people on the solution and helps people build beautiful visualizations with Chronograf. Within two years of production, the company has had no outage at all. Additionally, lots of libraries in many different languages exist, making data readily available for your programs. Hervieu recommends reading carefully what can be done and limitations of tags / measurements, as this is key to getting the most out of InfluxDB.