The Apache DolphinScheduler community has officially announced the release of Apache DolphinScheduler 3.4.2.
As an important maintenance release in the 3.4.x series, version 3.4.2 continues the community’s tradition of delivering high-quality iterations. This release introduces several user-facing enhancements, including the new Amazon EMR Serverless Task Plugin and improvements to the Monitoring Center. It also delivers extensive optimizations and refactoring across complement data processing, task plugin capabilities, security governance, permission control, and the underlying architecture.
From a release perspective, DolphinScheduler 3.4.2 includes more than one hundred merged pull requests covering feature enhancements, stability fixes, code governance, documentation improvements, and CI/CD optimizations. Contributors from across the global community participated in this release, further advancing DolphinScheduler as an enterprise-grade cloud-native workflow orchestration platform.
Amazon EMR Serverless Officially Joins the DolphinScheduler Ecosystem
As more organizations deploy data platforms on AWS, serverless architectures are becoming an increasingly important direction for big data computing.
To address this demand, the community introduced the Amazon EMR Serverless Task Plugin in this release (PR #18069, contributed by @norrishuang).
With this plugin, users can directly submit and manage Amazon EMR Serverless jobs from within DolphinScheduler, enabling unified orchestration and scheduling of Spark and other big data workloads without maintaining the underlying compute clusters. For data lakes, offline data warehouses, and AI-driven data processing scenarios, this translates into lower operational costs and higher resource utilization.
The addition of this capability further expands DolphinScheduler’s coverage within the AWS cloud-native ecosystem.
Enhanced Monitoring Center Delivers Greater Operational Visibility
In production environments, one of the most common questions for operations teams is understanding which workflows and tasks are currently running.
Previously, troubleshooting often required logging into servers and manually inspecting logs or process information. To improve this experience, the community enhanced the Monitoring Center through PR #18138 (contributed by @ruanwenjun).
After the upgrade, users can directly view workflows currently being executed by Active Masters and tasks currently running on Active Workers from the Monitoring UI. This provides a much clearer view of cluster activity and significantly improves troubleshooting efficiency.
For large-scale deployments, this enhancement further strengthens DolphinScheduler’s observability capabilities.
Complement Data Capabilities Further Improved
Complement Data (backfill processing) remains a core capability for offline data warehouses and modern data platforms.
In previous releases, dependency handling in certain backfill scenarios still had limitations. To address this issue, the community completed the implementation related to DSIP-95.
PR #18003, contributed by @det101, improves dependency task handling in complement data scenarios. After the upgrade, the system can correctly identify and calculate dependencies during historical data backfill operations, ensuring both accuracy and completeness throughout the backfill process.
This enhancement is particularly valuable for T+1 data warehouses, historical data recomputation, and offline metric correction scenarios.
Continued Enhancements to SQL, Flink, and Other Task Plugins
Task plugins continue to be one of the most actively evolving areas of DolphinScheduler.
For SQL Tasks, PR #18020 (contributed by @macdoor) introduces support for loading SQL statements directly from resource files while also supporting parameter placeholder substitution.
This enhancement allows complex SQL logic to be maintained in Resource Center files rather than embedded directly within task definitions. As a result, organizations can manage SQL assets more systematically while simplifying version control and collaborative development.
Within the Flink ecosystem, PR #17987 (also contributed by @macdoor) adds parameter substitution support to both Flink Task and FlinkStream Task plugins.
With this capability, users can dynamically generate task parameters at runtime, enabling workflows to adapt more easily to different environments and business scenarios. For organizations that manage multi-tenant configurations, dynamic date calculations, or environment switching, this feature significantly improves orchestration flexibility and operational efficiency.
In addition, the community further improved configuration flexibility for Flink tasks. PR #17909 (contributed by @leocook) exposes the Flink -sae parameter in the Web UI and sets its default value to disabled.
Through graphical configuration, users can control relevant runtime parameters without modifying underlying configuration files, further lowering the barrier to adopting Flink tasks within DolphinScheduler.
Continuous Improvements in Security and Permission Governance
Enterprise-grade scheduling platforms require robust security governance.
In this release, the community completed implementation work related to DSIP-37. PR #18119, contributed by @SbloodyS, introduces the ability to disable Jetty HTTP TRACE requests through configuration.
TRACE requests are rarely used in business scenarios but can potentially be exploited for information discovery. Disabling this capability helps reduce potential security risks and further strengthens system security.
Regarding datasource access control, PR #18073 (contributed by @njnu-seafish) introduces permission validation mechanisms for interfaces including connectionTest, getDatabases, getTables, and getTableColumns.
After the upgrade, unauthorized users can no longer retrieve related metadata information, further reinforcing security boundaries around data access.
Continuous Optimization and Refactoring
While new features are often the most visible improvements for users, consistent code standards, architectural refinements, and logic optimization are equally critical to the long-term sustainability of an open-source project.
In this release, the community continued optimizing and refactoring several core modules, including API, DAO, Datasource, Alert, Worker, and Storage. These improvements standardized implementation patterns, removed redundant logic, and further enhanced the maintainability and extensibility of the system.
Related pull requests include:
PR18103, PR18104, PR18105, PR18107, PR18111, PR18112, PR18113, PR18114, PR18115, PR18116, PR18117, and PR18118.
Meanwhile, @ruanwenjun continued advancing the Repository DAO refactoring initiative and the optimization of return object types.
Related pull requests include:
PR18226, PR18227, PR18228, PR18229, PR18230, PR18232, PR18233, PR18234, PR18236, PR18245, as well as multiple Repository refactoring pull requests from PR18250 through PR18263.
These improvements replace a large number of Map-based return structures with strongly typed objects, establishing a cleaner architectural foundation for future development.
Improved Stability Through Numerous Critical Fixes
In addition to feature enhancements, DolphinScheduler 3.4.2 resolves multiple critical issues encountered in production environments.
PR #18033 (contributed by @ruanwenjun) improves JDBC connection creation by switching to a Driver-based approach, resulting in more reliable database connectivity.
PR #18042 and PR #18044 (contributed by @njnu-seafish) fix parameter passing issues affecting Kubernetes Task and Zeppelin Task plugins.
PR #18155 (contributed by @SbloodyS) resolves an issue where date parameters were incorrectly propagated to sub-workflows during complement data execution.
PR #18146 (contributed by @SbloodyS) fixes a problem that could cause workflows using the CONTINUE strategy to remain in the RUNNING state indefinitely.
PR #18183 (contributed by @ruanwenjun) addresses a permission vulnerability that allowed users to delete task definitions from unauthorized projects.
PR #18212 and PR #18300 (contributed by @ruanwenjun) further strengthen permission validation for Workflow Trigger and Access Token-related operations.
In addition, the community resolved several other stability-related issues, including:
- Docker Compose deployment exceptions
- Helm Chart configuration problems
- JDBC Registry REMOVE event anomalies
- SQL License Header parsing issues
- Multiple defects affecting deployment reliability and runtime stability
Together, these fixes significantly improve the robustness of DolphinScheduler in enterprise production environments.
Continued Improvements to Documentation and Engineering Infrastructure
Beyond code-level enhancements, the community continues investing heavily in documentation and engineering excellence.
This release updates the FAQ documentation, improves parameter passing explanations, introduces AGENT.md and CLAUDE.md files, and fixes multiple documentation formatting and broken-link issues.
On the CI/CD front, the community restored Python end-to-end testing, improved unit test execution efficiency, optimized the Docker release process, and enhanced GitHub Code Owner management mechanisms.
Although these improvements may not appear directly in product feature lists, they serve as essential infrastructure that supports the long-term health and sustainability of the project.
Thanks to All Contributors
The successful release of Apache DolphinScheduler 3.4.2 would not have been possible without the collective efforts of contributors from around the world.
Special thanks go to Release Manager @ruanwenjun for his outstanding leadership and coordination throughout the release cycle.
Under his guidance, a total of 19 contributors successfully delivered this release:
@ruanwenjun
@SbloodyS
@det101
@njnu-seafish
@macdoor
@norrishuang
@leocook
@wcmolin
@asadjan4611
@HEEKDragonOne
@CloudExtreme
@Mrhs121
@hiSandog
@pjfanning
@sanjana2505006
@shaolei7788
@llphxd
@includetts
@shrihari7396
We would also like to thank everyone who contributed code, submitted issues, improved documentation, or provided valuable community feedback.
It is the continued dedication of every contributor that enables Apache DolphinScheduler to keep evolving and moving forward.
Download and Upgrade
Apache DolphinScheduler 3.4.2 is now officially available.
This release delivers substantial improvements across cloud-native capabilities, monitoring and observability, security governance, task plugins, and overall system stability. It provides a more reliable, efficient, and secure workflow orchestration experience for enterprise production environments.
We encourage all users to upgrade and explore the latest enhancements.
Download:
https://dolphinscheduler.apache.org/zh-cn/download/3.4.2
Release Notes:
https://github.com/apache/dolphinscheduler/releases/tag/3.4.2
Top comments (0)