Hello Community ,
We have developed several functionalities in the production enviroment that weren't in the previous version. There are process instances that are still running in the previous version but we would like to upgrade those instances so they can took all the new features that have been deployed in prod. Its my undestanting that for the new functionalites show up in the actual process, a new instance should be launched and the previous must be cancelled. Since we have hundreds of instancess running , its impossible to do it one by one.
Is there an automatic process or function in appian that could help us with this ? how do you approach this issue ??
thanks in advance!
Diego.
Discussion posts and replies are publicly visible
There is a plug-in that allows you to upgrade process instances to a new version: https://community.appian.com/b/appmarket/posts/process-management-services. However, it's a tricky procedure, as there's a lot of considerations based on what changes are made. There's a pretty good page in the documentation that provides Process Upgrade Guidance that I'd highly recommend reading before you consider an update.
Also some general guidance - in the future I'd recommend trying to set up your processes so they are short lived. This makes it less likely that a process upgrade is necessary and has lots of other benefits too (e.g. being more memory efficient). Here's a page in the playbook that goes through creating memory efficient models.
As Peter notes, this is tricky - there are a lot of considerations, when/where different data points are populated vs where your process instances are in their lifecycle, how drastic are the changes you are making, etc etc.
We've released thousands of enhancements over the past decade and never offer to the customer that running instances will be updated with the new logic. Only in one situation have we used the IFM Manager plugin, to resolve a bug that was deployed back on 16.x where a parent process passed a CDT to a sub process which contained a newer version, causing data to be cleared in legacy running instances.
Another read from the playbook is Creating Flexible Processes.
Oh good call - that playbook article is probably even more relevant than the one I posted!
Similar idea for both! ;)