Microservices Engine:XXL-JOB engine versions

New features

• Data Source Management: Associate a database or big data platform and run SQL tasks for data analytics.

• SQL tasks: Run SQL commands against an associated data source for data retrieval or analytics.

• IP address whitelist: Restrict access to specified IP addresses or CIDR blocks within your VPC.

• Workflow enhancements:

  • Import and export: Batch export and import workflows to synchronize them between clusters.

  • External dependency nodes: Fully compatible with DolphinScheduler depend nodes. Supports daily, weekly, and monthly dependency epochs.

  • Node disabling: Disable individual nodes within a workflow. Disabled nodes are automatically skipped during execution.

  • Reset scheduled time: Postpone nodes that have not yet started by manually resetting their scheduled time.

New features

• Workflow-based task orchestration: Orchestrate tasks using workflows with logical nodes, custom variables and outputs, upstream and downstream data transfer, dependency check rules, and O&M operations. Compatible with over 90% of common DolphinScheduler features.

• Calendar management: Create custom calendars and configure tasks to run on specific calendar schedules.

• Task start time: Schedule tasks to run at a future time or at a fixed time each day.

• Early completion alert: Receive an alert when a task finishes faster than the configured threshold.

Optimizations

• Increased the maximum timeout for HTTP tasks to 30 minutes.

Bug fixes

• Fixed an issue where logs of historical executions could not be viewed from the event query.

• Fixed a potential memory leak caused by broadcast sharding in engine version 2.3.0.

New features

• HTTP application executor registration: Register executors automatically through Kubernetes (K8s) service discovery, or manually enter domain names as executors.

• Regular application executor registration: Both automatic registration and manual entry are supported.

Bug fixes

• Fixed an issue where deleting a job while it was running could cause the job to get stuck.

• Fixed an issue where a job could get stuck if its pod IP address was taken by another application after deployment.

New features

• Job weight-based routing: Assign weights to jobs of different scales for better workload balancing.

• HTTP applications and HTTP jobs: Make HTTP calls by configuring domain names and interfaces, no SDK integration required.

• Sync jobs: Synchronize jobs between different instances.

Optimizations

• Saving a script job now shows historical version information to prevent overwriting previous versions.

• The execution list now includes an execution node field, making it easier to identify which node handled each schedule.

Bug fixes

• Fixed an issue where a job got stuck because the XXL-JOB client callback failed to deserialize.

• Fixed an issue where an old script was invoked during execution after the GLUE script content was updated.

New features

• Script version management for GLUE jobs. View and compare the version history of scripts.

• Tag-based routing policies. Control whether an untagged job can be scheduled on a tagged machine.

New features

• GLUE script tasks, including GLUE (Java), GLUE (Shell), GLUE (Python), GLUE (PHP), GLUE (Node.js), and GLUE (PowerShell).

• Go-based executor support. For more information, see Connection configuration.

• DingTalk alert cards support the <a> tag for opening alert links on mobile devices.

New features

• Dify flow tasks: Manage and schedule Dify workflows directly from XXL-JOB.

• Added duration and token consumption columns to the execution list.

Bug fixes

• Fixed an issue where fixed_delay jobs could stop scheduling if triggered frequently.

• Fixed an issue where a job status remained "failed" after a successful manual rerun.

• Fixed an accessToken authentication failure that occurred when the IP address of one application was reassigned to another after the executors for multiple applications restarted. This issue affects open source versions 2.1.x and earlier.

• Fixed an issue that caused jobs to get stuck when an executor process was forcibly killed and then immediately restarted with the same IP address and port. This issue affects all open source versions.

• Fixed an issue where business logs were not completely stored in the cloud.

New features

• Task orchestration using subtasks.

• Notifications for successful job results.

Optimizations

• Improved the notification style for successful and failed jobs.

New features

• View the stack for broadcast sharding jobs. Requires schedulerx3-plugin-xxljob 2.2.1 or later.

• If a shard fails, the error message is displayed and an alert is triggered.

Bug fixes

• Fixed an issue where manually stopping a shard in a broadcast sharding job could cause the job to get stuck.

New features

• Compatible with open source XXL-JOB 2.1.x versions.

• Job management alerts support configuring CloudMonitor contact groups.

• View the stack for running jobs. Requires schedulerx3-plugin-xxljob 2.2.1 or later.

• Graceful shutdown for clients. Requires schedulerx3-plugin-xxljob 2.2.1 or later.

Optimizations

• Disabling a job no longer automatically terminates running execution records. It only affects the next schedule, which is compatible with open source rules.

Bug fixes

• Fixed an issue where webhook alerts failed and the alert event notification showed an incorrect status.

New features

• fixed_delay time type: The next job runs N seconds after the previous one completes.

• Blocking policies compatible with open source strategies: standalone serial execution, discarding subsequent schedules, and overwriting previous schedules.

• Automatic termination on timeout: Stops the client's execution thread when a job times out.

• Three new properties on the job management page: Current Execution Status, Last Execution End Time, and Last Execution Status.

Bug fixes

• Fixed an issue where broadcast sharding jobs could end prematurely in rare cases.

• Fixed an issue where shard details for broadcast sharding jobs were sometimes displayed incorrectly.

• Fixed an issue where the "First" and "Last" routing policies were inconsistent.

New features

• Alert events.

• Tag-based routing.

• Automatic failover when a specified machine goes offline.

• Filter the executor list by tags.

• Alerts when no machines are available.

Bug fixes

• Fixed an issue where imported jobs could not be automatically scheduled.

• Fixed an expression parsing error for one_time jobs.

• Fixed an unauthorized access issue caused by an AppName and JobName mismatch.

• Fixed an issue where notification channel configurations did not take effect when creating a job.

• Fixed an issue where enabling log collection did not take effect when creating an application.

New features

XXL-JOB is now in public preview. This initial release is compatible with open source XXL-JOB clients 2.2.0 and later.