We are happy to announce that Stitch now supports native ETL into Snowflake. This integration enables users to load data from more than 60 data sources directly into a Snowflake data warehouse in minutes. Our customers have been giving us great feedback since we launched Snowflake integration into beta, and we’re excited to announce that it is generally available.
We built Stitch because modern cloud data warehouses are fundamentally different from the legacy, on-premises data warehouses of the past. These new technologies require a new model of ETL, and no data warehouse takes this model further than Snowflake. Stitch and Snowflake are a killer combination for your company’s modern analytics infrastructure because they align on separation of compute and storage, elastic scalability, support for rich data types, a cloud-built architecture, and a business model based on usage.
Separation of compute and storage
Snowflake has true separation of compute and storage, which enables far more flexibility than is possible with a traditional data warehouse. Unlike other data warehouses, when you need more storage, you can increase that without being forced to simultaneously increase compute since Snowflake keeps them separate. The ability to independently control compute and storage means that you you only pay for what you need.
Stitch enables Snowflake users to analyze all of their data by integrating with more than 60 data sources, including databases, SaaS tools, ad networks, and more. We also sponsor and integrate with the Singer open source ETL project, which makes it easy to get any additional or custom data sources into Snowflake. The combination of Stitch and Singer make sure that you leave no data behind, and Snowflake ensures that your data warehouse can handle anything you need to analyze.
Snowflake’s elastic scalability makes it fast and cost-effective to do in-database transformations. Traditional ETL — or extract, transform, and load — meant that data needed to be transformed prior to loading into the data warehouse. This made a lot of sense when your data warehouse was a physical piece of hardware sitting in your data center because provisioning additional resources might take weeks or months. With a data warehouse built for the cloud, you can scale it up or down, on demand. This allows users to load their raw data directly into the data warehouse immediately following extraction, with transformation happening afterwards, usually via SQL. This extract, load, transform workflow is known as ELT, and Stitch was built from the ground up to enable it.
Support for rich data types and SQL
Modern data sources such as web APIs generate data as JSON or XML, with complex nested data structures. Some data warehouses don’t support these types, and so we need to convert to a different data type or de-nest the data prior to loading. In Snowflake, these rich data types are natively supported and can be directly loaded and queried via SQL. This means customers get access to high-fidelity data from any data source, using the language they already know.
Both Stitch and Snowflake are cloud-native applications. You can be up and running in minutes, and there’s no maintenance to do, dials to turn, or hardware to manage.
We’ve also seen a growing trend of users who graduate to Snowflake after hitting the limits of other databases. Our destination switching feature makes it easy for existing Stitch users to take their existing integrations and start loading them into Snowflake right away.
Usage business model
Both Stitch and Snowflake are priced based on usage, which makes them economical for businesses of any size. Stitch offers an unlimited two-week free trial where you can load all of your historical data into Snowflake for free, and a free-forever tier for small data volumes where you can load up to five million records per month.
Start your free trial of Stitch today and load data into Snowflake in minutes.