August 2022 Summaries

Timescale Cloud is being used by many customers for mission-critical applications that require close monitoring to prevent service interruptions. To monitor these services, users can integrate Timescale Cloud with Amazon CloudWatch, which is especially useful if they are already using CloudWatch to monitor other AWS infrastructure components. The integration allows users to monitor memory, CPU, and storage metrics in Timescale Cloud, providing a convenient way to run TimeSeriesDB databases in production. To set up the integration, users need to create an exporter in Timescale Cloud that sends data to Amazon CloudWatch, and then attach this exporter to their desired service. The process involves creating a CloudWatch user with necessary credentials, exporting metrics from Timescale Cloud, and configuring the export settings. Once the integration is set up, users can access their service metrics via CloudWatch, providing real-time visibility into their database performance.

Timescale Cloud Tips discusses the importance of chunk size in hypertables, which are the foundation of TimescaleDB. The size of chunks affects performance and can impact various features such as compression, continuous aggregates, and data retention policies. Understanding priorities, characteristics of data, and server resources is crucial to fine-tuning the chunk time interval setting. A good starting point for chunk time interval is 14 days, which allows for high compression ratios, minimal storage footprint, and efficient data retention. Continuous aggregates require a separate chunk time interval setting to accommodate data retention, and this should be modified immediately after creating the aggregate. Optimizing the chunk time interval parameter can significantly impact TimescaleDB performance and efficiency, and understanding data, server resources, and application requirements is essential for fine-tuning this setting.

To improve your ingest rate in Timescale Cloud, set proper expectations by aiming for 50-100k rows per second per ingest process. Ensure you have enough resources on both the client and server sides, including CPU and network bandwidth. Inserting data in batches with multiple rows per INSERT statement can also optimize your ingest rate. By following these tips, you can maximize your row inserts per second and improve your overall performance in Timescale Cloud.

This installment of the "Community Member Spotlight" series highlights the work of Shane Steidley, director of software at Density, and Brock Friedrich, software engineer. They explain how they are using TimescaleDB to query data from IoT devices to help companies with substantial real estate footprints make data-driven decisions to optimize their real estate usage while reducing carbon emissions. Density is a company that builds technology to help companies understand how people use physical space. They use sensors and analytics platforms to provide comprehensive insights into how spaces are used, allowing companies to compare a space's performance against its intended purpose, other spaces, or portfolio benchmarks. The team uses TimescaleDB to store counts of people in spaces over time and to perform continuous aggregates, which have been a game-changer in their development process. They find that using TimescaleDB has made development faster and more efficient, allowing them to offload decision logic and complexity around time zone handling and bucketing to the database. The team is now moving towards ETL-type processing out of the database but still utilizing continuous aggregates for rolling up high-resolution data to lower resolutions. They recommend resources such as the Timescale blog and advise developers to consider real-time use cases and historical use cases separately when designing a scalable database or strategy around an analytic workload.

When customers first learn about high availability in Timescale Cloud, they often express concerns about giving their data to a third-party vendor. However, hosted databases like Timescale Cloud can provide a reliable and scalable solution, with AWS handling the management and reliability of underlying hardware. The platform's decoupled compute and storage architecture allows for rapid recovery from compute failures, which can be automatically triggered without user intervention. Additionally, replicas can be enabled to ensure near-zero downtime and data loss in case of database failures. Backups are also maintained, but their use is not the primary method of data protection, as they can result in longer downtime compared to replication. By understanding how Timescale Cloud handles high availability, users can make informed decisions about their workload's requirements and choose a solution that meets their needs.