Timescale announced the availability of Timescale Cloud, a fully managed version of their time series database on Azure, GCP, and AWS. It provides time series analysis functions, scale up and down, visualization integration with tools like Grafana and Tableau, and data encryption.
TimescaleDB is an open source extension to PostgreSQL that is optimized for manipulating time series data. Timescale, the company behind it, announced the general availability of the managed solution on various public cloud providers. Unlike most other time series databases, TimescaleDB is built on top of a relational database -- PostgreSQL -- as an extension. It uses an abstraction called hypertable to hide partitioning across multiple dimensions like time and space. Each hypertable is split into "chunks", and each chunk corresponds to a specific time interval. PostgreSQL supports restore, point-in-time recovery and replication, and the same applies to the time series data stored in it using Timescale.
Aiven had released support for TimescaleDB as a PostgreSQL extension on GCP, AWS and DigitalOcean last year. So what is different about this release? "Aiven supports the open-source version of TimescaleDB", says Ajay Kulkarni, co-founder/CEO, TimescaleDB, whereas Timescale Cloud "includes enterprise features like interpolation and data retention policies."
The managed solution does not autoscale; the user has to choose the correct hardware configuration for their workload as VM instances. Migrating is possible between different instance types, with a short downtime. Diana Hsieh, director of product management at TimescaleDB, explains that they "spin up a separate instance that matches the type that you want to migrate to, restore a backup and stream the WAL logs. Then we redirect (to the new instance)."
With time series databases growing in popularity and scale of usage, many cloud providers offer them as managed solutions. AWS and GCP both have managed PostgreSQL - but neither of them support TimescaleDB yet, whereas Azure does. AWS also has its own managed time series database called Timestream, which is still in preview mode. Timescale Cloud's pricing depends on the cloud provider and the chosen hardware configuration.