I have a problem, if I have a process and I have a timer on a node of an integration, when the timer will start running, when the process starts or when the node runs?
Greetings and thank you!
Discussion posts and replies are publicly visible
The documentation: "The delay or schedule added to an activity or a process flow only begins once the node or event has been activated."
What if we want you to consider milliseconds? because we have put this, so that it goes by the exception in case in which 1 seconds pass but we see that the integration takes 1,096 and instead of going by the exception it is continuing through the normal flow
now () + 1 / (24 * 60 * 60)
Greetings and thank you
I don't understand what you're asking. As a principle you cannot assume an integration is going to respond either at all or within a reasonable timeframe. So you can put a timer on the node. Typically you'd want to wait (for a process that has a real User waiting for it to respond) something like 3-5 seconds (note: the business requirements would usually say what this value should be). You'd never want to express this in milliseconds, which is why I don't understand why you'd want to define a value to this level of precision.If you meant 'seconds then note that an Integration object now has an explicit timeout attribute available:
...which you can define in seconds.
We have the problem because we have set a timer in the integration of 1 second (to test later it will be higher), it is assumed that if the integration takes more than 1 second to run, it must exit by exception and only continue the flow, but I see in the process node that is taking 1,097 seconds and instead of going through the exception goes through the normal flow.
Greetings and thank you very much!
I've just configured an example as follows:
...and verified that it does indeed follow the exception in 1 second.
Where are you seeing the value of 1,097 seconds?
In the proccess model tab, this is a case of another execution but it does the same, instead of going for the exception it is coming out of the ok flow.
thank you very much!
I suspect the value you're seeing (that is > 1 second, but only a little bit over) represents the total time spent in the task, including start-up and tear-down, so don't thinbk this is significant.
The fact that it's following the Ok flow is more perplexing. Can you share a screen-shot of a completed instance? And also of where you've configured the Exception?
the flow advances and does not go by exception
Do you have a timeout set in your Integration object?
In the version we have that is 19.1 I have not seen anywhere that you can configure the timeout on the integration object.
The configuration is in the node, where we call integration in the process.
Ok, that's a new feature so obviously not available in 19.1.
I suspect what is happening is that the service IS responding faster than the timer value you've set, but appearing to be slower because of the overhead of starting and ending the node. If you test the integration object natively what is the response time there?
Also, try to set the timer on the node to 0.5s to see if that forces the flow down the exception path.
If we execute the integration as it takes the following:
I have put in the timer configuration:
Even so, it continues through the normal flow and does not go through the exception
© 2020 Appian. All rights reserved.