I have a DataLink that can takes 1-5 minutes depending of the number of subprocesses to create. It calls an Epicor function which then calls the ECM API to create one or multiple documents. Each calls takes about 30-45 seconds so the 3 minutes limit seems broken for a longer example. It still creates all the subprocesses but it’s annoying to get an exception in ECM because the user won’t really think about fixing it since it’s an asynchronous task
I found another post where they updated the web.config with that value but it’s only for AP2 and SO integrations it seems
<add key="ErpTimeoutSeconds" value="300" />
<!--affects precreate and create in AP2 and SO integrations; default is 180; place this in Astria.Client.WindowsServices.AutomationService.exe.config for client-based integration-->
If I run the function via the Epicor API (Swagger page , v2), everything is OK. It’s really only the message on the ECM side which is causing problem in my workflow
The ECM support guys are usually pretty good and I suspect you’ll need to get to them to figure out how to change (find out which) timeout settings - or have them help figure out why it’s cancelling in general.
Other option would be to figure out how to make the Function trigger asynchronously and release the ECM workflow datalink call more quickly. Perhaps the data link does a UD table row insert, and the function is called by a directive on the UD table row insert?
Wondering if you tried other versions v1 or different datalinks to do the same call, sometimes a REST call works in REST and doesn’t work in ECM, I usually try to use a different call since timeouts and other calls can be difficult to diagnose in ECM.
I have reported a similar issue, one of our PO has 60 lines and when ECM is doing the matching, it will become a task is cancelled. Case number : CS0004146994
I have experienced this with a number of customers and many end up submitting a ticket to Epicor to have the issue resolved. Based on the amount of anecdotal instances there appears to be for this, it would seem that the ECM development team should look at a better limit to the timeout to avoid this issue.
In some cases, the Auto-Match (the typical culprit of timeouts) can be amended to reduce the timeframe of potential matches. This has resolved the issue in some cases, but not all. It may also have the inverse result for some invoices compared to others. As a result, this may require multiple instances of the Auto-Match action to be designed for certain vendors to filter results by date while others don’t need that.
Yeah, it seems like it’s not well documented enough. Support just got back to me asking how many lines of invoices I’m creating which means that they don’t understand my issue at all since I’m not creating invoices at all
For now, I’ll try to put a variable in the return message of my DataLink and if my variable is not empty, I’ll send an email to myself in another action. So, 95% of the time, the workflow will advance correctly and the other 5%, when it triggers an exception, I’ll receive an email to fix it myself fast
