[DebeziumIO] Implement startOffset & offset persistence for Kafka (#28248)

[tkaymak]

Implement withStartOffset for Debezium and external offset persistence to handle restarts. (Addresses #28248)

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a crucial startOffset capability to DebeziumIO, allowing users to define a specific point for Debezium connectors to begin or resume data consumption. This enhancement is vital for building resilient data pipelines that can recover from failures or planned shutdowns without losing data or re-processing unnecessary events. The implementation spans across the Go, Java, and Python SDKs, providing a unified experience for managing change data capture streams.

Highlights

• DebeziumIO Start Offset Implementation: Introduced a startOffset feature for DebeziumIO in both Go and Java SDKs, allowing users to specify a point from which the Debezium connector should resume consuming changes.
• Cross-language Compatibility: The startOffset functionality is integrated across the Go, Java, and Python SDKs, ensuring consistent behavior for DebeziumIO users regardless of their chosen language.
• Offset Parsing and Handling: The Java SDK now includes logic to parse startOffset values provided as key-value strings, automatically converting numeric parts to Long where appropriate, facilitating flexible offset specification.
• Improved Restart Capabilities: This feature significantly enhances the robustness of Debezium-based pipelines by enabling them to gracefully restart from a known good state, preventing reprocessing of already handled data.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• sdks/go/pkg/beam/io/xlang/debeziumio/debezium.go
  • Added StartOffset field to the readFromDebeziumSchema struct to hold offset configuration.
  • Implemented a new StartOffset function that serves as a read option to configure the starting offset for the Debezium connector.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumIO.java
  • Added getStartOffset and setStartOffset abstract methods to the Read and Builder classes, respectively, to manage the new offset property.
  • Introduced the withStartOffset builder method, allowing users to specify a map of key-value pairs representing the desired starting offset for the Debezium connector, complete with detailed Javadoc.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumTransformRegistrar.java
  • Imported HashMap for map manipulation.
  • Added a startOffset field to the Configuration class and its corresponding setter method.
  • Implemented logic within buildExternal to parse the list of startOffset strings into a Map<String, Object>, converting numeric values to Long where applicable, before applying it to the readTransform.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/KafkaSourceConsumerFn.java
  • Modified the getInitialRestriction method to pass the newly introduced startOffset from the spec to the OffsetHolder.
• sdks/java/io/debezium/src/test/java/org/apache/beam/io/debezium/DebeziumIOTest.java
  • Added static imports for assertEquals and assertNull.
  • Imported ImmutableMap for easier map creation in tests.
  • Included new test cases to verify that withStartOffset correctly stores the offset, handles null offsets, throws exceptions for invalid null input, and that getInitialRestriction properly utilizes the provided start offset.
• sdks/python/apache_beam/io/debezium.py
  • Imported Dict and Union from typing.
  • Extended ReadFromDebeziumSchema to include an optional start_offset field, defined as a list of strings.
  • Updated the ReadFromDebezium constructor to accept a start_offset parameter and included comprehensive documentation for its usage and format.
  • Passed the start_offset parameter to the ReadFromDebeziumSchema during initialization.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[github-actions]

Assigning reviewers:

R: @tvalentyn for label python.
R: @ahmedabu98 for label java.
R: @jrmccluskey for label go.

Note: If you would like to opt out of this review, comment assign to next reviewer.

Available commands:

• stop reviewer notifications - opt out of the automated review tooling
• remind me after tests pass - tag the comment author after tests pass
• waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

[tvalentyn]

Hi @an2x @johnjcasey @Dippatel98 - could you please help find a reviewer for this? Thanks

[tkaymak]

@tvalentyn maybe @Abacn can have a look (?)

[Abacn]

Thanks, this change allows user to provide an offset that was obtained elsewhere. It now supports to read from an offset, yet not completely resolve #28248, as restarting the pipeline that have set the offset would still start from the same offset.

I had a (yet materialized) idea of restartable offset. Basically we can provide an "OffsetRetainer" interface to 1. read offset on pipeline startup; 2. write offset on checkpointing, and expose as a configuration for the IO. And we can then provide a built-in FileSystem based OffsetRetainer for user to use.

When pipeline is running, it continuously write committed offset to external location (gcs, kafka, etc) of choice; when pipeline starts or restarts, it tries to read from the same location at first.

If you prefer to go for current change for now, let me take a closer look for this PR. Or we can work on a more concrete solution.

[tkaymak]

@Abacn thank you for comment. I think your solution sounds like a much better vision, and we should aim for that.

[tkaymak]

@gemini-code-assist summary

[gemini-code-assist]

Summary of Changes

This pull request significantly improves the resilience and usability of DebeziumIO by introducing robust offset management features. It allows users to either explicitly define a starting offset for Debezium connectors or leverage an automatic file system-based mechanism to persist and restore offsets, ensuring seamless pipeline restarts and fault tolerance.

Highlights

• Enhanced Offset Management: Introduced startOffset and OffsetStoragePath capabilities for DebeziumIO in both Go and Java SDKs, allowing users to specify a point from which the Debezium connector should resume consuming changes or automatically manage offsets via a file system.
• Automatic Offset Persistence: Implemented an OffsetRetainer interface and a FileSystemOffsetRetainer class in Java, enabling automatic saving and loading of Debezium connector offsets to any Beam-compatible filesystem (e.g., GCS, S3) after each successful checkpoint.
• Cross-Language Consistency: Ensured that the new offset management features are consistently exposed and configurable across the Go, Java, and Python SDKs, providing a unified experience for DebeziumIO users.
• Improved Pipeline Resilience: Significantly enhanced the robustness of Debezium-based pipelines by allowing them to gracefully restart from a known good state, preventing data loss or reprocessing of already handled events.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• sdks/go/pkg/beam/io/xlang/debeziumio/debezium.go
  • Added StartOffset and OffsetStoragePath fields to the readFromDebeziumSchema struct.
  • Implemented new StartOffset and OffsetStoragePath functions as read options for configuring the Debezium connector.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumIO.java
  • Added abstract getStartOffset, getOffsetRetainer, setStartOffset, and setOffsetRetainer methods to the Read and Builder classes.
  • Introduced withStartOffset and withOffsetRetainer builder methods with detailed Javadoc for configuring offset management.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumTransformRegistrar.java
  • Imported HashMap for map manipulation.
  • Added startOffset and offsetStoragePath fields to the Configuration class and their corresponding setters.
  • Implemented logic to parse startOffset strings into a Map<String, Object> and to configure OffsetRetainer based on offsetStoragePath.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/FileSystemOffsetRetainer.java
  • Added a new class FileSystemOffsetRetainer to persist and load Debezium connector offsets as JSON files using Beam's FileSystems.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/KafkaSourceConsumerFn.java
  • Modified getInitialRestriction to prioritize loading offsets from an OffsetRetainer or falling back to an explicit startOffset.
  • Added logic to save the committed offset using the OffsetRetainer after each successful task commit.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/OffsetRetainer.java
  • Added a new OffsetRetainer interface for defining strategies to persist and restore Debezium connector offsets.
• sdks/java/io/debezium/src/test/java/org/apache/beam/io/debezium/DebeziumIOTest.java
  • Added new test cases to verify the functionality of withStartOffset and withOffsetRetainer, including priority handling and null checks.
• sdks/python/apache_beam/io/debezium.py
  • Extended ReadFromDebeziumSchema to include optional start_offset and offset_storage_path fields.
  • Updated the ReadFromDebezium constructor to accept and document start_offset and offset_storage_path parameters.

Activity

• A summary of changes was provided by gemini-code-assist[bot].
• Reviewers were assigned by github-actions[bot].
• tvalentyn requested assistance in finding a reviewer.
• tkaymak suggested Abacn as a potential reviewer.
• Abacn provided feedback, suggesting a more comprehensive OffsetRetainer interface for restartable offsets.
• tkaymak agreed with Abacn's vision for a better solution.
• tkaymak requested a new summary.

[tkaymak]

I would appreciate a second look and opinion on this @Abacn 🤝

[Abacn]

Sorry for the delay (back from OOO).

[tkaymak]

Sorry for the delay (back from OOO).

Absolutely no need to apologize! :)
I have adjusted and rebased according to your comments - very appreciated!

@gemini-code-assist summary

[gemini-code-assist]

Summary of Changes

This pull request introduces robust offset management capabilities to DebeziumIO, significantly enhancing its resilience and usability. It provides users with the flexibility to either explicitly define a starting offset for Debezium connectors or leverage an automatic file system-based mechanism to persist and restore offsets. This ensures seamless pipeline restarts and fault tolerance by allowing pipelines to resume data consumption from the last successfully processed point, preventing data loss and unnecessary reprocessing.

Highlights

• Comprehensive Offset Management: Introduced both explicit startOffset and automatic OffsetRetainer mechanisms for DebeziumIO in Go, Java, and Python SDKs, allowing flexible control over data consumption resumption.
• Automatic File-System Based Persistence: Implemented a FileSystemOffsetRetainer in Java, enabling automatic saving and loading of Debezium connector offsets to any Beam-compatible filesystem (e.g., GCS, S3) after each successful checkpoint, ensuring fault tolerance with atomic writes.
• Cross-Language API Consistency: Ensured that the new offset management features are consistently exposed and configurable across the Go, Java, and Python SDKs, providing a unified experience for DebeziumIO users.
• Robust Pipeline Restart Capabilities: Significantly enhanced the resilience of Debezium-based pipelines by allowing them to gracefully restart from a known good state, preventing data loss or reprocessing of already handled events, with OffsetRetainer taking precedence over startOffset.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog

• sdks/go/pkg/beam/io/xlang/debeziumio/debezium.go
  • Added StartOffset and OffsetStoragePath fields to the readFromDebeziumSchema struct.
  • Implemented StartOffset and OffsetStoragePath functions as read options for configuring the Debezium connector.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumIO.java
  • Added abstract methods getStartOffset, getOffsetRetainer, setStartOffset, and setOffsetRetainer to the Read and Builder classes.
  • Introduced withStartOffset and withOffsetRetainer builder methods with detailed Javadoc for configuring offset management.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/DebeziumTransformRegistrar.java
  • Imported HashMap for map manipulation.
  • Added startOffset and offsetStoragePath fields to the Configuration class and their corresponding setters.
  • Implemented logic to parse startOffset strings into a Map<String, Object> and to configure OffsetRetainer based on offsetStoragePath.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/FileSystemOffsetRetainer.java
  • Added a new class FileSystemOffsetRetainer to persist and load Debezium connector offsets as JSON files using Beam's FileSystems.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/KafkaSourceConsumerFn.java
  • Modified getInitialRestriction to prioritize loading offsets from an OffsetRetainer or falling back to an explicit startOffset.
  • Added logic to save the committed offset using the OffsetRetainer after each successful task commit.
• sdks/java/io/debezium/src/main/java/org/apache/beam/io/debezium/OffsetRetainer.java
  • Added a new OffsetRetainer interface for defining strategies to persist and restore Debezium connector offsets.
• sdks/java/io/debezium/src/test/java/org/apache/beam/io/debezium/DebeziumIOTest.java
  • Added new test cases to verify the functionality of withStartOffset and withOffsetRetainer, including priority handling and null checks.
• sdks/java/io/debezium/src/test/java/org/apache/beam/io/debezium/FileSystemOffsetRetainerTest.java
  • Added unit tests for FileSystemOffsetRetainer, covering offset loading, saving, error handling, and serialization.
• sdks/python/apache_beam/io/debezium.py
  • Extended ReadFromDebeziumSchema to include optional start_offset and offset_storage_path fields.
  • Updated the ReadFromDebezium constructor to accept and document start_offset and offset_storage_path parameters.

Activity

• An initial summary of changes was provided by gemini-code-assist[bot].
• Reviewers were assigned by github-actions[bot] to cover Python, Java, and Go changes.
• tvalentyn requested assistance in finding a reviewer for the PR.
• tkaymak suggested @Abacn as a potential reviewer.
• Abacn provided feedback, suggesting a more comprehensive OffsetRetainer interface for restartable offsets and atomic writes.
• tkaymak agreed with Abacn's vision for a better solution.
• tkaymak requested a new summary from gemini-code-assist[bot].
• gemini-code-assist[bot] provided an updated summary reflecting the new offset management features.
• tkaymak requested a second look from @Abacn.
• tkaymak adjusted and rebased the PR based on Abacn's review comments.
• Abacn provided review comments on FileSystemOffsetRetainer.java regarding surfacing exceptions for unreadable files, using temporary files for atomic writes, skipping writes if the offset is unchanged, and using static factory methods. tkaymak acknowledged the suggestion for static methods.

[Abacn]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces a significant feature to DebeziumIO by implementing startOffset and offset persistence capabilities, which are crucial for creating resumable pipelines. The changes are well-integrated across the Go, Java, and Python SDKs. The implementation includes a new OffsetRetainer interface in Java, with a FileSystemOffsetRetainer for durable storage, and updates the cross-language configuration accordingly. The code is well-documented and includes comprehensive tests. I have one suggestion to improve the robustness of the cross-language configuration handling in Java to prevent potential exceptions with malformed input.

[Abacn]

LGTM pending AI comment

Tested locally the FileSystemOffsetRetainer functionality

Also consider adding an item for https://github.com/apache/beam/blob/master/CHANGES.md

[tkaymak]

@Abacn Thank you! Implemented the review comment and updated CHANGES.md

[Abacn]

Thank you!

[tkaymak]

@Abacn pleasure!
(I can't merge as I don't have the rights)