Everactive is an industrial IoT (IIoT) company that combines battery-free, self-powered sensors and powerful cloud analytics to provide end-to-end IIoT solutions to their customers. Everactive’s batteryless sensors enable them to monitor factory devices and collect data in a more economical way, which means that their industrial data streams are more detailed and cover more equipment than competitors. Everactive chips are built in-house, designed for harsh settings, and can operate indefinitely from low levels of energy harvested from heat or light. Everactive’s customers include Fortune 500 manufacturers in many industries, including process manufacturing and chemical processing.
Everactive experienced three main problems with PostgreSQL and non-time-series databases: major performance and indexing problems, a lack of queries specific to their types of data, and the lack of a quick onboarding process for engineers as the company scaled. Before TimescaleDB, Everactive stored metadata in PostgreSQL and sensor data in OpenTSDB. Over time, OpenTSDB became increasingly slow and brittle—it had become normal for OpenTSDB to crash multiple times per week from users asking for too much data at once.
Everactive chose TimescaleDB because it let them double down on PostgreSQL, which they loved using for metadata about their packet streams. They looked briefly at competitors like InfluxDB but stopped considering them once it was clear TimescaleDB would exceed their needs. TimescaleDB was faster than their previous OpenTSDB system. More importantly, nobody has ever crashed it.
Everactive ingests data from thousands of sensors into TimescaleDB, then surfaces it to their customers through dashboards, charts, and automated alerts. Everactive shows customers their data through a web-based dashboard and also has internal applications to help their in-house domain experts review and label customer data to improve their automated failure detection.
One not very carefully benchmarked but huge performance increase they’ve seen? They have a frontend view that requires the last data point from all sensors. In OpenTSDB, it required nearly 10 minutes to load (due to hard-to-fix tail latencies in HBase). With their first TimescaleDB deployment? It only took around seven seconds. Further improvements to their schema and access patterns have brought these queries into the sub-second range.
, Data Science Staff Engineer
Data Science Staff Engineer