March 2021 Summaries

Creating a schema in TimescaleDB differs from creating it in PostgreSQL, with two key differences: including a time field in unique keys and not having foreign keys within hypertables. The optimal chunk size should fit into ¼ of the available memory to support heavy operational workloads. Creating a separate unit key can speed up query times for certain queries, but not all. Hardware requirements depend on dataset size and memory usage, with recommendations available through Timescale's Slack channel or benchmarking blog posts. Migrating from PostgreSQL (non-hosted) to Timescale Cloud is relatively easy, using tools like pg_dump and psql, and can be done with minimal downtime in the future.

We're excited to share two new open-source tools from the Promscale team, announcing that TimescaleDB 2.0 is officially generally available, sharing our favorite tools, tips, and advice for doing more with your time-series data, and continuing to release new features, create new documentation and tutorials, and host virtual sessions to help developers do amazing things with their data. We're on a mission to teach the world about time-series data, supporting and growing communities around the world by sharing educational resources as broadly as possible. Our goal is to make observability more accessible to more developers, removing the need to compromise on flexibility, ease of use, or cost. We've released new open-source tools such as Tobs for deploying a full observability suite for Kubernetes in two minutes and Prom-migrator for migrating Prometheus data to and from popular remote-storage systems. We're also sharing our favorite new pieces in our biweekly newsletter and hosting virtual sessions to help developers learn more about time-series data and how to do amazing things with it.