Datadrift is an open-source metric observability framework that helps data teams deliver trusted and reliable metrics.

Data monitoring tools fail by focusing on static tests (eg. null, unique, expected values) and metadata monitoring (eg. column-level).

Datadrift monitors your metrics, sends alerts when anomalies are detected and automates root cause analysis.
Data teams detect and solve data issues faster with Datadrift's row-level monitoring & troubleshooting.

pip install driftdb

Here is a quick demo. For a step-by-step guide on the dbt installation, see the docs.

Install the monitor in your pipeline.

>>> from driftdb.connectors import LocalConnector
>>> LocalConnector().snapshot_table(table_dataframe=dataframe, table_name="revenue")

For a step-by-step guide on the python installation, see the docs.

We are in development and we would love to do the installation with you. Fill the form on our website so we can do a 15min demo. If the tool solves your problem then the installation requires 30min.

Get full visibility into metrics variation and pro-actively detect data quality issues. Become aware of unknown unknowns with metric drift custom alerting.

Operationalize your monitoring and solve your underlying data quality issue with lineage drill-down to understand the root cause of the problem.

Give visibility to data analysts and data consumers with shared explanation of metric variation.

We are in the early days of Datadrift. Just open a new issue to tell us more about it and see how we could help!

We 💚 contributions big and small. In priority order (although everything is appreciated) with the most helpful first:

• Star this repo to help us get visibility and build awesome open-source tools
• Join our Discord server to be part of our thriving community
• Open an issue to share your idea or a bug you might have spotted
• Become a Design Partner to co-built a product you & users love

Track planning on Github Projects and help us prioritising by upvoting or creating issues.