REFramework questions/use case

Help Something Else
1

Hi! I’m learning the REFramework and trying to implement in some business cases. Reframework called my attention because of its cappability to handling errors, but when i look to my process some questions appears, for example:

I have a process with 3 steps:

1- input the order in ERP
2- Alocate the order
3- Separate the order

Once one of these steps are done they can not execute again for the same order. Imagine if i had a system exception in the second step, it can not retry to process the transaction from the first step.

Should i have one robot for each step? For example: the first robot input the orders and populate a queue with the orders generated, the second get these queue itens, alocate and populate another queue, the thrid uses this last queue to separate the orders

Should i use something like a checkpoint? For example: after pass the first step some boolean variable is set to True

Should i do not use retry transaction?

2

Hi @guilherme.dias,

Each step can feed other tail. In this way, you can run the end-to-end process using the trigger method.

With status control, you can control transactions in a single process. By filling out the output field in the queue, you can continue from the step left in the records with errors.

Regards,
MY

1 Like
3

Hi @guilherme.dias

There would always be some cases where you encounter a system exception for reasons beyond your control. In most cases they can be resolved by retrying the transaction. However, it does get tricky for cases like inputting information to generate a case or work order number. Depending your specific scenario, you can try any of these options that fit your needs:

  • You can look into adding a custom progress while the processing the item. You can check the progress and only execute steps that haven’t been completed yet.

  • Check in the system if the work has been completed before executing those steps. E.g. You can search for order in ERP and only create if it isn’t.

  • Lastly, you can split up the work like you suggested. Depending on the volume, you can have each robot perform a specific task, or simply use separate queues for each task. Robot will create a new queue item in Queue2 after it successfully executes a transaction in Queue1 and so on. Though this option will become more difficult to manage, troubleshoot, provide artifacts to auditors or even generate reports for business.

1 Like
4

Thank you @grkiran2 . Add a custom progress might work for my process! I’ll test it right now!

5

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.