Hello,
I am currently using REFramework to execute process from Orchestrator, it has around 200 transactions to execute. Currently I am running separate workflow to add these items to queue and then i am starting execution from Orchestrator to process these items.
Can someone share best practices for this, as I don’t want to use 2 process to accomplish task?
The REFramework model suggests that this is the best way to handle the data. The Dispatcher process loads the data, and the Performer process processes it.
However, this doesn’t work for everybody. If you need this to be one process, you could load the data in the part of the REFramework that only runs once when the process starts. This is good for smaller data loads that require only a single bot.
Take a look at the Enhanced Reframework Enhanced REFrameWork - RPA Component | UiPath Marketplace.
Your dispatcher would be a service in this case.
If that framework seems to complicated, you could create a project which is Ref in Ref.
Or instead of add queue item, use add transaction item (https://docs.uipath.com/activities/docs/add-transaction-item) which adds the item in the queue and sets its progress to “In Progress” and then immediately process the transaction.
The case of a combined dispatcher/processor, where the transactions are immediately processed, how are failed transactions recommended to be handled as no solely processing process exists? I’d imagine the dispatcher part of the process would need an “add item only if it doesn’t already exist” part, which I think means a definable unique ID must be used?
Is the recommendation to use two processes generally?