3 Replies Latest reply on Oct 8, 2008 12:26 PM by Ronald van Kuijk

    Modeling a retry

    James Cook Newbie

      I have an action that authorizes a credit card. The communication to the payment processor might fail for one reason or another. Because of this, if there is a failure, I would like to retry the call after a 2 minute delay. If that fails I will try again after 5 minutes. If communication fails a third time, I will transition to a manual process.

      Would I model this authorization process as three separate nodes (auth1, auth2, and auth3) or is there a better approach using one node and schedule or timers?

        • 1. Re: Modeling a retry
          Ronald van Kuijk Master

          There are so many ways to model this, none of which probably is wrong. 2 nodes is an option, one with timers, one with a real task (so it is visible in the process) Three nodes is even more visible, but that is up to you...

          • 2. Re: Modeling a retry
            James Cook Newbie

            Great, thanks. I just wanted to make sure there wasn't a way to do it with a single node.

            • 3. Re: Modeling a retry
              Ronald van Kuijk Master

              We'll there might be, just not sure if it will work, therefor I did not suggest it.

              You could try making a task node with a task that is not created when the node is entered. Then put the two timers on them but the second timer action should be slightly different. It should create the task instead of leaving the node. But... if a task-node has no tasks, the token automatically continues. This is not what you want, so you should set signalling to false... this is just from the top of my head so things might be missing, but if feels fabricated...