This action plug-in is a tool to balance request between a pool of resources that have specific capacities. It provides a way to pause a workflow until the enough resource becomes available and route the request to the resource the less taxed.
# => 0.0.1 first tracked version (requires ManagedResource object introduced in 2.0.0)
# => 0.0.2 fixes race condition if multiple triggers are used in conjunction with a single queue
# => 0.1.0 added keep trigger on going mode
# => 0.1.1 2015-03-05 ML - New icon
# => 0.1.2 2015-06-10 ADM - Use the queue_item_description as allocation label unless differently specified or not available
# => 0.1.3 2015-12-11 ML - Fix issue preventing use of RUNTIME queue
# => 0.1.4 2015-11-26 ADM - Fixed possibility to specify resource capacity at run time
# => 0.1.5 2016-01-10 ADM - New category for resource management
# => 0.1.6 2016-03-22 PM - Selection of the RUN TIME value for queue does not persist in the dropdown although saved correctly in the database
# => 0.1.7 2016-07-29 ADM - Perform a first check for resource availability ASAP
# => 0.1.8 2016-09-21 ML - Frees the allocated resource in case a proper lock on queued item was not obtained
# => 0.1.9 2017-05-19 HF - Fix hold timeout ignored
The load manager plugin allows to manage resources in a mutual exclusion way, similarly to the resource manager. But unlike the resource manager, you can associate a number of slots to a resource.
For example, you could have 2 transcoders, each of them capable of handling 3 max concurrent jobs. With the resource manager, you can handle this case by defining groups of resources in the resource list (such as IP1_1, IP1_2, IP1_3, IP2_1, IP2_2, IP2_3) and then by extracting the resource id from the composed name. With the load manager, you can declare a capacity (i.e. a number of slots) to allocate to each resource.
A portlet has also been designed to display the load manager resource utilization: managed load utilization