This topic describes the release notes for Realtime Compute for Apache Flink and provides links to relevant references. The release notes provide the major updates and bug fixes in Realtime Compute for Apache Flink in the version that was released on December 16, 2021.
Platform updates
New features
Inspect deployment logs and analyze performance
To use this feature, enable log archiving in your log template, or turn on Allow Log Archives in the Logging section on the Configuration tab for the deployment.
The platform now surfaces three categories of deployment logs directly in the Realtime Compute for Apache Flink console: startup logs, operational logs, and historical logs. For running deployments, you can also analyze JobManager and TaskManager performance on the Flame Graph, Memory, and Threads tabs.
For more information, see View the logs of a historical job and View the startup logs and operational logs of a deployment. To analyze performance, see Monitor deployment performance.
Create Hive, Hologres, and MySQL catalogs from the console
Hive, Hologres, and MySQL catalogs can now be created directly from the console UI.
Check state compatibility before restoring a job
When you select a historical state to restore a job, the system checks state compatibility and reports the potential impact on your running business.
Define reusable variables in SQL
General variables — including AccessKey pairs — can be configured and referenced across SQL statements. For more information, see Manage variables and keys.
Optimizations
Chain analysis results and operator-level resource configuration added to fine-grained SQL
-
Chain analysis results are now displayed in the console.
-
Operator-level resource configuration is now supported.
Deployment name added to alert messages
Alert messages now include the deployment name, making it faster to identify which deployment triggered an alert.
Catalog creation UI redesigned
The UI for creating catalogs has been redesigned for a smoother experience.
Ververica Runtime (VVR) updates
VVR 3.0.7
-
Deployment exploration is now supported.
-
Known issues in Apache Flink 1.12 are fixed.
VVR 4.0.11
Flink Change Data Capture (CDC) commercial features
Two CDC capabilities are now generally available:
-
Schema evolution: Tables can evolve their schema.
-
Database synchronization SQL semantics: Flink SQL now supports the semantics needed for database synchronization workflows.
For the SQL statements that power these features, see CREATE TABLE AS statement and CREATE DATABASE AS statement. For an end-to-end walkthrough, see Ingest data into data warehouses in real time.
Gemini state backend supports OSS
The Gemini state backend can now store state data in Object Storage Service (OSS).
Other improvements
-
Deployment resource management is optimized.
-
The MySQL CDC DataStream connector is improved. For more information, see MySQL CDC DataStream connector.
-
An exception in the Gemini state backend is fixed.
-
Known issues in Apache Flink 1.13 are fixed.