How Trebellar Halved Storage Costs While Unlocking Real-Time Insights With PostgreSQL

In case you didn’t hear, at Timescale, we think you should use PostgreSQL, a database developers know and love, for everything, any use case, across any industry. Throughout the year, we’ve continued to make improvements to TimescaleDB and our managed PostgreSQL offering, making it easier to use PostgreSQL for Everything. Most recently, in August, we recapped a set of performance improvements, and in September, a set of releases dedicated to improving developer experience.

This time, we’re focused on helping you get to production faster with PostgreSQL and extending the AI toolkit so you can build better search and retrieval-augmented generated (RAG) applications with PostgreSQL. 

As we look forward to seeing these releases being used in the wild—from pgai Vectorizer to the SQL Assistant—fewer things motivate us more than learning how we impacted your application development for the better. In this blog post, we catch up with the real estate management platform Trebellar to see how they’re cutting storage costs in half while ingesting 10 million data points daily and supporting real-time insights for their users.

Trebellar: Providing Real-Time Insights From IoT Data

Trebellar provides a platform for real estate management, streamlining data collection and analysis for large commercial properties. They transform data from IoT sensors (such as temperature, humidity, occupancy, and air quality) to help building managers optimize their operations. 

By integrating diverse data streams into a centralized platform, Trebellar helps clients in retail, hospitality, storage, and more make informed decisions about their space utilization and building efficiency.

The Trebellar team built a pipeline that ingests and normalizes sensor data from any source to monitor and predict building efficiency, eliminating silos in building management data. There are three layers within that pipeline:

1. The data layer: data collection and normalization
2. The insights layer: machine learning to enable predictive analytics 
3. The action layer: actions and solutions based on generated insights

Performance Challenges and Improvements

Before Timescale, Trebellar was facing challenges in processing large volumes of time-series data generated from their customers' IoT devices. They needed a solution that could handle the high frequency of data inputs, normalize it efficiently, and support real-time insights for their platform's machine learning models. The complexity of managing data across different devices, formats, and locations made it difficult to provide actionable insights in a timely manner.

The Trebellar engineering team had always liked PostgreSQL as a battle-tested, gold-standard open-source database option. They selected Timescale not only for the power of TimescaleDB but also for the seamless integration with PostgreSQL. For Trebellar, it wasn’t just the power of the platform but the quality of the documentation, community, content, and more.

With TimescaleDB, Trebellar significantly improved their ability to manage and query large datasets. With the Timescale automation framework and features like hypertables, time-bucketing, and compression, Timescale enabled them to downsample and compress data effectively, reducing storage costs by 50 percent. 

"We capture 10 million points, 10 million rows a day. We need to downsample that after a month and compress it. I can do that so seamlessly with essentially five lines of code. If Timescale didn't exist, perhaps we would have tried to just do something directly with PostgreSQL, but that would have resulted in much worse performance.” David, Co-Founder, Trebellar 

Now, Trebellar can provide real-time analytics and insights to their customers, streamlining decision-making for building management and optimizing energy usage, occupancy, and air quality monitoring.

Watch the full story:

Trebellar is one of many companies building better applications with PostgreSQL on Timescale, but there are many more we want to celebrate, like our friends at Nocodelytics, SolarNetwork, and Sentinel Marine Solutions. 

In their words:

“I’m using Timescale because it’s the same as PostgreSQL but magically faster." Florian Herrengt, Co-Founder at Nocodelytics

“We already used PostgreSQL and Timescale was appealing to us because it provided a better way to manage the table partitioning and adapt it as our usage grew.” Matt Magoffin, Technical Director, SolarNetwork

“Timescale allows us to scale our services without introducing completely new technologies to the mix. As PostgreSQL users, Timescale adds very little maintenance overhead compared to learning and maintaining a brand new database system.”
Pedro Kostelec, CTO at Sentinel Marine Solutions

Stay tuned throughout the week for most Timescale updates and product news! If you’d like to try Timescale free for 30 days, you can sign up here. And if you’d like to be notified of future releases, be sure to sign up for the Timescale newsletter!