August 2020 Summaries

The TimescaleDB community has been active this month, with exciting product releases, including improved Managed Service for TimescaleDB, now available in 75+ regions and 2K configurations. The team also shared various tutorials, events, and how-tos to help developers master time-series data, such as using PostgreSQL to speed up Grafana and visualizing timeshifts. Community members were recognized with shoutouts, including a successful cryptobot built using TimescaleDB. The newsletter rounded up new content, including a session replay on improving PostgreSQL INSERT performance and a blog post from Timescale CTO Mike Freedman. Additionally, the team shared product updates, such as Timescale Cloud now including CPU, memory, and storage metrics reporting, and a Prometheus Adapter beta release. The community was also invited to participate in remote-friendly events, including Office Hours and a Reddit AMA.

The problem with Grafana's drill-down functionality is that it can be slow for non-aggregated, fine-grained data. To solve this, the solution uses PostgreSQL's UNION ALL function to allow "auto-switching" between aggregations depending on the time interval selected. This enables users to create graphs that automatically switch between different aggregations of their data, allowing them to drill down into their metrics as quickly and efficiently as possible while saving time and CPU resources. By using continuous aggregates in PostgreSQL, users can speed up their dashboards even further by not directly querying the hypertable that houses raw data.

Felipe Queis, a senior full-stack engineer, uses TimescaleDB to power his crypto trading bot, which has achieved impressive results, including a success rate of 61.5% and a cumulative gross result of approximately 487%. He was initially drawn to cryptocurrencies after a friend's servers were infected with ransomware, and he started creating a simple Moving Average Convergence Divergence (MACD) crossover bot. However, he soon realized that his bot needed significant improvements and started working on it in his spare time. Felipe's project has evolved into a sophisticated system with technical indicators, sentiment analysis powered by machine learning, and a capital management system. He uses TimescaleDB to store and process the large amounts of data generated by his bot, which includes real-time aggregations, high ingestion rates, and efficient storage and compression. With TimescaleDB, Felipe's query response time is in the milliseconds, even with huge datasets. He has successfully integrated TimescaleDB into his existing technology stack, including Node.js, TensorFlow, and Cote, without adding significant maintenance overhead. Felipe's experience showcases the power of time-series data to fuel real-world decisions, and he has become an evangelist for the use of Time-scaleDB in the developer community.

We're always releasing new features for TimescaleDB, a fully-managed relational database service for time-series data, including the Public Preview of Timescale Cloud, which offers a fully-managed cloud service with independent scaling of compute and storage. The team has also released TimescaleDB 1.7.2, which includes updates to continuous aggregates, downsampling, and compression. Additionally, they've created step-by-step guides for connecting with various programming languages, such as Python, Ruby, and Node.js, to get started with Time-series analysis using TimescaleDB. The community has also been sharing their experiences with TimescaleDB, including a fintech startup that chose the database for its scalability and a radio transmission analytics project using WsprDaemon. The team is proud to now offer two fully-managed relational database services for time-series data and will continue to release new content and events, including virtual sessions and community spotlights.