Skip to content

OpenMetadata 0.12.0-Release
Compare
Choose a tag to compare
@Vj-L Vj-L released this 08 Sep 19:16
· 14 commits to 0.12.0 since this release
0.12.0-release
56afd08
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

0.12.0 Release - Sept 7th 2022 🎉

You can read the Release Blog here
or watch an awesome video showing the new features!






Team Hierarchy

Prior releases supported a flat hierarchy of just Teams and Users. In 0.12, support has been added for the entire organizational hierarchy with Business Unit, Division, Department, and Groups. An organization from small to very large can now be modeled in OpenMetadata with this feature.

Roles and Policies

Access Control functionality has been revamped to support many use cases that were not possible before. Previously, a Role contained a single Policy, which consisted of simple Rules to Allow/Not Allow. The advanced rule configuration in the 0.12 release allows users to build more expressive rules using conditions.

  • A Role is a collection of Policies. Roles can be assigned to users or teams where all the users in the team inherit the team roles.
  • A Policy is a collection of Rules. A Policy can be reused as it can be part of a Role or can be directly assigned to Teams.
  • A Rule is defined by a set of Resources, a set of Operations, an Effect to either Deny or Allow the operation, and a condition written as SpEL expression to add additional conditions based on metadata attributes. Examples of conditions — isOwner(), noOwner() && !matchTags('PII').

Data Quality and Data Profiler

OpenMetadata began support for Data Quality in the 0.10 release, and support was added for publishing Great Expectations results in the 0.11 release. Our goal with OpenMetadata is to define metadata standards for all things data and in this release, we are standardizing Tests and Data Quality metadata. Data Quality Tests can be expressed in JSON schema and now these tests can be added dynamically using the Test Definitions API. We have also added a custom SQL data quality test that allows you to write your data quality tests using SQL statements.

An interactive dashboard helps to visualize and explore the data from the Data Profiler. You can explore how your data is changing over time, and identify data drifts using this dashboard. You can also see how data quality is changing by looking at how tests are doing over time. What is even better is, that you can explore this at both the table level or drill down to each column level going back up to 60 days.

The UI supports the detailed exploration of data quality tests, and users can drill down for the details of the test results present in a time series fashion. Tests can be added easily from the Profiler tab in the UI, both at the Table and Column levels. The UI provides a one-glance update on the metrics with a summary of data quality at the Table and Column levels.

Announcements

Informing users about upcoming changes to the data is a big challenge. In most organizations, a team sends an email well in advance about the change. But no one reads/tracks them and finally, when the change is done, many users are unprepared to handle it.

With Announcements, you can now inform your entire team of all the upcoming events and changes, such as deprecation, deletion, or schema changes. These announcements can be scheduled with a start date and an end date. All the users following your data are not only notified in Activity Feeds but a banner is also shown on the data asset details page for users to discover (or be reminded of) the announcement.

Activity Feed Notifications

In 0.12, we’ve also streamlined the Notifications menu with two separate tabs for Tasks and Mentions, that’ll display only the recent notifications. You can always navigate to your User Profile page to view more activities.

Slack & Microsoft Teams integration

Users can get timely updates about the metadata change events for all entities through APIs using webhooks. The webhook integration with Slack has been further improved in this release.

OpenMetadata also supports webhook integration to Microsoft Teams, just as it supports Slack. Users can choose to receive notifications for only the required entities by using event filters based on when an entity is created, updated, or deleted.

Tasks

In the 0.11 release, a request to add or update descriptions for data assets could be converted to a Task. In the 0.12 release, Tasks can be created based on requests to create or update tags. Also, a glossary term approval workflow can be converted to a Task.

Secret Management Store Interface

In 0.12, we have completely revamped how that secret is stored, accessed, and by whom; by introducing a Secrets Manager Interface to communicate with any Key Management Store. The KMS will mediate between any OpenMetadata internal requirement and sensitive information. That way, users can choose to use the underlying database as KMS, or any external system. The OpenMetadata community has already added support for AWS Key Management Service and AWS SSM.

Connectors

New connectors are an essential part of every release in OpenMetadata. We are introducing four new connectors in this release:

  • Redpanda is a Kafka API-compatible streaming data platform for developers that unifies historical and real-time data. OpenMetadata now supports Redpanda as a Messaging service, which allows users to document its topics and schemas. Refer to the Redpanda documentation for more info.
  • Dagster is a new-generation Python-based orchestrator that’s designed for developing and maintaining data assets, such as tables, data sets, machine learning models, and reports. It has been added as part of OpenMetadata’s pipeline connectors. Read more from the Dagster documentation.
  • Fivetran delivers ready-to-use connectors that automatically adapt as schemas and APIs change, ensuring consistent, reliable access to data. It has been added as a pipeline service. For more information, refer to the Fivetran documentation.
  • Apache NiFi automates the flow of data between systems. OpenMetadata now supports a NiFi connector as the third new pipeline service on this release.

Lineage

We’ve enhanced the performance of workflows by having a separate workflow for Lineage and Usage. By using two workflows for computing specific pieces of information, we can effectively filter down the queries to extract lineage.

During table usage ingestion, the tables retrieved successfully will be cached, so that there is no need to repeat the same calls multiple times as many queries would be referencing the same tables.
Usage queries have been optimized.
A result limit has been added to Usage queries.

Global Settings

The OpenMetadata Settings dropdown menu has been transformed into a single, centralized Settings page for added convenience in viewing all the available options. The Global Settings comprises setting options for Team Members, Access based on Roles and Policies, Services, Data Quality, Collaboration, Custom Attributes, and Integrations for webhooks and bots. Admins can view or update settings for various services like Slack, MS Teams, Webhooks, etc from the Global Settings page.

UI/UX Improvements

The major UI UX improvements have been done around Roles and Policies and a Global Settings page. Quite a lot of tweaks have been made to the UI to improve the UX.

When creating a new user or when a user is registering for the first time, the dropdown menu for Teams now displays an option to ‘Show All’ teams. Previously, we supported the display of only the first 10 teams. An option has also been provided to search and filter.
UI improvements have been made on the Schema, Service, and Database details pages.
Manage Tab has been replaced with the manage button on the UI.

New Contributors

Full Changelog: 0.11.0-release...0.12.0-release

Contributors

k-popov, rameeshm, and 13 other contributors
Assets 5