State change boundaries

Continuing on the industrial IT theme, an organization managing a large estate of devices needs to identify operating trends and rhythms across its devices to control costs and enable proactive measures, such as predictive maintenance. In this example, you will discover how to build an aggregation pipeline to identify patterns for when devices are in use or idle.

Note

For this example, you require MongoDB version 5.0 or above. This is because you will use time-series collections, the $setWindowFields stage, and the $shift operator introduced in version 5.0.

Scenario

You are monitoring various industrial devices (e.g., heaters and fans) contained in the business locations of your clients. You want to understand the typical patterns of when these devices are on and off to help you optimize for sustainability by reducing energy costs and their carbon footprint. The source database contains periodic readings for every device, capturing whether each is...