Snowflake expects the updates and observability features to simplify devops and accelerate development.
Cloud-based data warehouse company Snowflake has added a new development interface for machine learning pipelines to its platform, among other updates for developers announced at the ongoing Snowflake Summit.
These updates includes the new interface, called Snowflake Notebooks; the addition of a Pandas API; new observability features; and the integration of the company’s Native App Framework with Snowpark Container Services.
Snowflake Notebooks, which is currently in public preview, is natively integrated to all parts of the Snowflake platform, including Snowpark ML, Streamlit, and Cortex. It’s a single development interface for Python, SQL, and Markdown that developers can use to experiment and iterate on their machine learning (ML) pipelines, harness AI-powered editing features, and simplify data engineering workflows, the company said.
Meanwhile, a Snowpark Pandas API will enable Python developers to work with the familiar syntax of the Pandas open-source Python library for loading, manipulating, aligning, merging, and visualizing data tables directly in Python. The Pandas API is currently in public preview.
Updates to devops tools
In order to help with devops, Snowflake said that it has added features such as Database Change Management to define the desired state of data pipelines with infrastructure-as-code principles rather than scripting complex workflows line by line, and Git integration to enable development collaboration across teams and streamlines deployments across different environments. Both are in public preview.
Other devops updates include the Python API and Snowflake CLI moving to general availability soon.
Snowflake adds observability capabilities via Trail
The company also added new observability features in the form of Snowflake Trail, which provides visibility into data quality, pipelines, and applications, enabling developers to monitor, troubleshoot, and optimize their workflows. It is built with OpenTelemetry standards so developers can integrate with popular observability and alert platforms including Datadog, Grafana, Metaplane, PagerDuty, and Slack, among others.
Additionally, Snowflake is providing built-in telemetry signals for Snowpark and Snowpark Container Services, enabling users to easily diagnose and debug errors using metrics, logs, and distributed tracing—without having to manually set up agents or transfer data.
In order to help build applications faster, the company said that it was integrating the Native App Framework into Snowpark Container Services.
“The integration enables enterprises to extend the breadth and variety of applications they build in the AI Data Cloud using configurable GPU and CPU instances to fit a range of use cases spanning computer vision automation, geospatial data analysis, and ML applications for enterprises,” the company said.
