Optimizing resources with the affinity attribute

​Efficient resource allocation and task synchronization are essential for workflow management. The affinity attribute ensures that tasks are run on the same workstation dynamically, improving performance and minimizing delays. This feature is especially valuable in environments using pool workstations.

What is Affinity?
The affinity attribute allocates tasks to run on the same workstation. This is set at the task level or the workflow level. It ensures multiple tasks dynamically share the same workstation, enhancing performance and simplifying resource management.
 
Key Benefits
Dynamic Resource Allocation: aligns tasks on the same dynamically chosen workstation.
Improved Workflow Efficiency: reduces overhead and enhances synchronization.
Resource Optimization: maximizes the use of workstation capabilities.
 
Affinity Requirements
Task Workflow Alignment: the tasks must belong to the same workflow.
Workstation Compatibility: the dynamic agents of the tasks must match. The workstation pools must either match or share members.
If these conditions are not met, the first task waits for 10 minutes (default) before failing.
 
Customization Options
Wait Time: set the UNO_WAITING_JOB_EXPIRE_TIMEOUT parameter in the “values.yaml” file to adjust the default 10-minute limit.
Monitoring Frequency: modify the UNO_WAITING_JOB_EXPIRE_FREQUENCY parameter to adjust the default 60+ second monitoring interval.
 
Priority in Definitions
Affinity can be added to:
Task Definitions: directly included during task configuration.
Workflow Definitions: applied for broader control.
Note: workflow-level definitions override task-level configurations.
 
Implementation Example:
Configure the definition as follows to run the tasks TASKDEF_1 and TASK2, run on workstation /WS_AGT_0 in the delta_pool:-

Conclusion
​
By leveraging the affinity attribute, tasks are run on the same workstation, reducing overhead and optimizing resource usage. Implement these strategies to enhance workflow reliability and guarantee optimal performances.

---

​Peddi Pranav - Technical Lead at HCL Software

Pranav became a member of HCL Software in August 2022 and currently contributes to the UnO Backend team. He specializes in technologies like Java, Spring Boot, RESTful Web Services, Hibernate, Microservices and MongoDB. Outside of work, he enjoys embarking on adventures to new destinations and tackling puzzles as a hobby.