[
https://issues.apache.org/jira/browse/UNOMI-878?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Serge Huber updated UNOMI-878:
------------------------------
Description:
The SchedulerService provides a robust, cluster-aware task scheduling system
with several valuable features that improve reliability, scalability, and
developer experience:
Key Features:
1. Integrated Persistence Layer:
- Seamless integration with Unomi's PersistenceService for task storage
- Automatic task state persistence and recovery
- Support for both persistent and in-memory task storage
- Efficient task querying and filtering capabilities
2. Advanced Cluster Support:
- Node-specific task execution control with `executorNode` configuration
- Ability for nodes to opt-out of task execution
- Automatic task distribution across cluster nodes
- Task execution isolation through node-specific locking
- Built-in crash recovery for failed nodes
- Support for running tasks on all nodes or specific nodes
3. Enhanced Developer Experience:
- Fluent Builder API for intuitive task creation
- Simple recurring task creation for common use cases
- Comprehensive task lifecycle management
- Rich task monitoring capabilities
4. Robust Task Management:
- Automatic task recovery from node crashes
- Configurable retry policies
- Task resumption from checkpoints
- Automatic task purging with configurable TTL
- Support for both one-shot and recurring tasks
Example Usage (New Builder Pattern):
{code:java}
// Create a cluster-wide persistent
taskschedulerService.newTask("dataSync")
.withPeriod(1, TimeUnit.HOURS)
.withSimpleExecutor(() -> syncData())
.schedule();
// Create a node-specific memory
taskschedulerService.newTask("localCache")
.nonPersistent()
.withPeriod(5, TimeUnit.MINUTES)
.withSimpleExecutor(() -> cleanCache())
.schedule(); {code}
Technical Benefits:
1. Improved System Reliability:
- Automatic crash recovery prevents task loss
- Persistent storage ensures task survival across restarts
- Cluster-aware execution prevents task conflicts
2. Enhanced Scalability:
- Flexible node participation in task execution
- Efficient task distribution across cluster
- Support for both cluster-wide and node-specific tasks
3. Better Resource Management:
- Option to run memory-only tasks for better performance
- Automatic task cleanup to prevent storage bloat
- Configurable thread pools for task execution
4. Operational Excellence:
- Comprehensive task status tracking
- Built-in monitoring capabilities
- Automatic task recovery and cleanup
Migration Impact:
- Fully backward compatible
- New features are opt-in
- Existing task definitions continue to work as before
was:
The SchedulerService provides a robust, cluster-aware task scheduling system
with several valuable features that improve reliability, scalability, and
developer experience:
Key Features:
1. Integrated Persistence Layer:
- Seamless integration with Unomi's PersistenceService for task storage
- Automatic task state persistence and recovery
- Support for both persistent and in-memory task storage
- Efficient task querying and filtering capabilities
2. Advanced Cluster Support:
- Node-specific task execution control with `executorNode` configuration
- Ability for nodes to opt-out of task execution
- Automatic task distribution across cluster nodes
- Task execution isolation through node-specific locking
- Built-in crash recovery for failed nodes
- Support for running tasks on all nodes or specific nodes
3. Enhanced Developer Experience:
- Fluent Builder API for intuitive task creation
- Simple recurring task creation for common use cases
- Comprehensive task lifecycle management
- Rich task monitoring capabilities
4. Robust Task Management:
- Automatic task recovery from node crashes
- Configurable retry policies
- Task resumption from checkpoints
- Automatic task purging with configurable TTL
- Support for both one-shot and recurring tasks
Example Usage (New Builder Pattern):
{code:java}
// Create a cluster-wide persistent
taskschedulerService.newTask("dataSync").withPeriod(1,
TimeUnit.HOURS).withSimpleExecutor(() -> syncData()).schedule();// Create a
node-specific memory
taskschedulerService.newTask("localCache").nonPersistent().withPeriod(5,
TimeUnit.MINUTES).withSimpleExecutor(() -> cleanCache()).schedule(); {code}
Technical Benefits:
1. Improved System Reliability:
- Automatic crash recovery prevents task loss
- Persistent storage ensures task survival across restarts
- Cluster-aware execution prevents task conflicts
2. Enhanced Scalability:
- Flexible node participation in task execution
- Efficient task distribution across cluster
- Support for both cluster-wide and node-specific tasks
3. Better Resource Management:
- Option to run memory-only tasks for better performance
- Automatic task cleanup to prevent storage bloat
- Configurable thread pools for task execution
4. Operational Excellence:
- Comprehensive task status tracking
- Built-in monitoring capabilities
- Automatic task recovery and cleanup
Migration Impact:
- Fully backward compatible
- New features are opt-in
- Existing task definitions continue to work as before
> Enhanced Cluster-Aware Task Scheduling Service with Improved Developer
> Experience and Persistence Integration
> -------------------------------------------------------------------------------------------------------------
>
> Key: UNOMI-878
> URL: https://issues.apache.org/jira/browse/UNOMI-878
> Project: Apache Unomi
> Issue Type: Bug
> Components: unomi(-core)
> Affects Versions: unomi-3.0.0
> Reporter: Serge Huber
> Priority: Major
> Fix For: unomi-3.0.0
>
>
> The SchedulerService provides a robust, cluster-aware task scheduling system
> with several valuable features that improve reliability, scalability, and
> developer experience:
> Key Features:
> 1. Integrated Persistence Layer:
> - Seamless integration with Unomi's PersistenceService for task storage
> - Automatic task state persistence and recovery
> - Support for both persistent and in-memory task storage
> - Efficient task querying and filtering capabilities
> 2. Advanced Cluster Support:
> - Node-specific task execution control with `executorNode` configuration
> - Ability for nodes to opt-out of task execution
> - Automatic task distribution across cluster nodes
> - Task execution isolation through node-specific locking
> - Built-in crash recovery for failed nodes
> - Support for running tasks on all nodes or specific nodes
> 3. Enhanced Developer Experience:
> - Fluent Builder API for intuitive task creation
> - Simple recurring task creation for common use cases
> - Comprehensive task lifecycle management
> - Rich task monitoring capabilities
> 4. Robust Task Management:
> - Automatic task recovery from node crashes
> - Configurable retry policies
> - Task resumption from checkpoints
> - Automatic task purging with configurable TTL
> - Support for both one-shot and recurring tasks
> Example Usage (New Builder Pattern):
> {code:java}
> // Create a cluster-wide persistent
> taskschedulerService.newTask("dataSync")
> .withPeriod(1, TimeUnit.HOURS)
> .withSimpleExecutor(() -> syncData())
> .schedule();
> // Create a node-specific memory
> taskschedulerService.newTask("localCache")
> .nonPersistent()
> .withPeriod(5, TimeUnit.MINUTES)
> .withSimpleExecutor(() -> cleanCache())
> .schedule(); {code}
> Technical Benefits:
> 1. Improved System Reliability:
> - Automatic crash recovery prevents task loss
> - Persistent storage ensures task survival across restarts
> - Cluster-aware execution prevents task conflicts
> 2. Enhanced Scalability:
> - Flexible node participation in task execution
> - Efficient task distribution across cluster
> - Support for both cluster-wide and node-specific tasks
> 3. Better Resource Management:
> - Option to run memory-only tasks for better performance
> - Automatic task cleanup to prevent storage bloat
> - Configurable thread pools for task execution
> 4. Operational Excellence:
> - Comprehensive task status tracking
> - Built-in monitoring capabilities
> - Automatic task recovery and cleanup
> Migration Impact:
> - Fully backward compatible
> - New features are opt-in
> - Existing task definitions continue to work as before
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
