✨ Feature Highlights

Fivetran integration

Sifflet now integrates with Fivetran, allowing you to gain unparalleled visibility into your data's end-to-end journey, from source to destination.

With this release, the Sifflet lineage graph no longer starts at the data platform level. Instead, you now have access to two types of nodes that represent your data’s journey before it even reaches the data platform:

• Fivetran connector nodes: These clickable nodes represent your Fivetran connectors, allowing you to get all the useful information about a given connector with just one click. This information includes the connector’s source and destination, its current status, sync frequency, and the timestamp of the latest sync. Accompanied by the Sifflet metadata that you’re accustomed to (owners, description, and last metadata refresh), this panel gives you, at a glance, all the necessary information about a given connector.
• Source asset nodes: For every Fivetran connector, Sifflet will create nodes for the source assets from which Fivetran extracts data, giving you complete visibility on your data’s journey. These nodes, which allow you to get a truly end-to-end lineage within Sifflet, represent one of the key differentiators of our Fivetran integration.

By integrating your Fivetran account with Sifflet, you'll also be able to swiftly detect connector issues (in the Sifflet data catalog and via the Fivetran connector nodes in the lineage) and easily get a bird’s-eye view of your Fivetran connectors, no matter where you encounter them in the lineage.

To get started with the integration, refer to the dedicated documentation page. For additional information about the exciting features of this integration, check out the official release announcement.

🛠 Fixes

• Fixed a bug that occasionally prevented Tableau and Looker assets from being connected to their upstream nodes in the lineage.

App version: v269

✨ Feature Highlights

Improved Experience on Incidents List Page

The incidents list page got revamped to improve the overall incident search user experience by making it consistent with the rest of the Sifflet application. By default, only Open and In progress incidents now show on the list, making it easier for you to focus on data quality issues that require immediate attention.

Read more about incidents

App version: v267

✨ Feature Highlights

Improved Control of Automatic Incident Creation

There are some monitors for which you don't necessarily want to create a Data Incident. You can now control Sifflet's automatic creation of an incident via the bulk edit from the Monitor list page.

This feature works also with dbt monitors.

In addition to the bulk edition of monitors, we've also made this toggle available at the dbt datasource level.

🛠 Fixes

• Fixed Technical Error in some instances of Duplicate and Null monitors with Group By
• Fixed Field Duplicate and Completeness monitors reporting incorrect results for empty datasets

App version: v265

🛠 Fixes

• Fixed a bug that caused Big Query partitioning not to be taken into account for some monitors.

App version: v257

✨ Feature Highlights

Bulk Edit Monitors Severity

Assigning the appropriate severity to a monitor is instrumental to ensure data quality issues are tackled with the right level of urgency by your teams. You can now bulk edit your monitors' severity, making it easier for you to update your monitor suite to ensure it is set to the proper criticality level.

Read more about monitors

🛠 Fixes

• Fixed a bug that caused user-entered dashboard descriptions to be overwritten during source refreshes.

App version: v256

✨ Feature Highlights

Credentials Now Redacted by Default

The content of credentials is now by default obfuscated when you edit them to simplify compliance with security teams requirements.

To read more about credentials

App version: v255

✨ Feature Highlights

Improved Classification and Documentation Recommendations

Sifflet AI Assistant was updated in order to:

• Provide better columns classification tags and documentation recommendations by leveraging OpenAI GPT-4o
• Optimise its performance and reduce its footprint on your environment when performing AI suggestions

Read more about AI Suggestions

Simplified Compliance With 20+ New Classification Tags

Sifflet AI Assistant recognizes 20+ new classifiers such as Email Address, Credit Card Number, IP Address or Login Details. Sifflet AI Assistant detects columns to which these classifiers apply and suggests them as classification tags, helping you categorise your data and consequently simplify data governance and compliance with regulation.

Read more about AI Suggestions and classification tags

Sifflet Insights Now Features Ownerships for Simpler Collaboration

Sifflet Insights (Sifflet browser extension) now surfaces information about assets' owners. This means data stakeholders can now know who to contact in case of issue or question on the dashboard they are looking at without even leaving their BI tool.

Read more about Sifflet Insights

New "Creator" Monitor Filter

You can now locate monitors you and people on your team created more easily thanks to Monitors list's new Creator filter category.

Read more about monitors

App version: v253

✨ Feature Highlights

Document and Classify Your Data in One-Click

You can now enrich your Data Catalog in one-click thanks to Sifflet AI Assistant new capability. Clicking the Generate metadata button on your data assets now indeed generates:

• Columns classification suggestions to help teams categorize Tables & Views assets' columns in a more accurate and comprehensive manner.
• Descriptions recommendations to enrich data assets documentation and ensure the content of datasets' columns is clear to all data stakeholders.

Learn more about AI Suggestions

Select Multiple Tables in Custom SQL Monitors

Available for Custom Metrics and Custom SQL

You can now select multiple tables when creating a Custom SQL monitor, this will allow the monitor to be linked to multiple tables at once but what it also truly unlocks is the power of AI to generate even more advanced Monitors, joining tables automatically and on demand !

Sifflet's AI Assistant will attempt to join tables together with the context of the tables' schemas and column descriptions! (if you identify your column as foreign key in the description that will be taken into account! )

🛠 Fixes

• [Data Catalog] Fixed a bug that was preventing some search parameters (search terms, search on fields selection, page, and number of items per page) from being maintained during a Data Catalog navigation.

App version: v252

✨ Feature Highlights

Freshness Improvements

Sifflet has in the past primarily been doing Freshness based on the distribution of the data, checking that data points with new dates were added. However this does not cover all freshness usecases, such as reloading a dimension table without a date field. In those cases what I want to check if if the data reload occurred within the usual time range !

New Template: Freshness (Update Time Gap)

This template can be applied to any table and will use the Table Metadata available to build up a history of typical time gaps between updates.

Due to the lack of history available, this monitor needs training time to be effective so you will need to wait a few data update cycles for a nice looking graph !

Supported Databases:

• Snowflake
• Bigquery
• Databricks
• Oracle
• Mysql

🛠 Fixes

• Fixed a bug that caused recurring data source update failures in some large environments.

Release: New monitors wizard, Rolling aggregation for monitors with dynamic thresholds

✨ Feature Highlights

New monitors wizard

We have improved the monitor's creation experience on the UI by introducing a redesigned wizard. This new user experience streamlines navigation within the monitor's templates, providing users with more precise guidance on setting parameters tailored to their specific use cases. We will continue improving this experience with more to come in the coming releases.

Rolling aggregation for monitors with dynamic thresholds

Introducing a new addition to our monitoring capabilities, detect anomalies using rolling aggregation alongside dynamic thresholds. Every monitor run will aggregate data based on user-defined rolling time intervals. For instance, users can set up monitors to run daily and detect anomalies in the total number of orders over the past seven days using dynamic thresholds. This feature is compatible with the following monitor templates:

• Metric (Dynamic)
• Duplicate Count (Dynamic)
• Duplicate Percentage (Dynamic)
• Null Count (Dynamic)
• Null Percentage (Dynamic)

App version: v246