You are viewing limited content. For full access, please sign in.

Discussion

Discussion

Improved handling of terminated instances?

posted on November 4, 2019 Show version history

Hi team,

I wanted to find out whether there are any plans to improve the way that Laserfiche Forms business processes will handle 'terminations' in upcoming releases?

Specifically, if you have a significantly complex case management/business process it only takes a single improperly configured inclusive/exclusive/parallel gateway to result in a total loss of an instance with no ability to recover it. This sort of irrecoverable data loss due to error is definitely unusual for enterprise products, so I'm curious to know whether or not it is even on the roadmap? If it isn't, I'd certainly like to raise my hand to ask for it to be given some consideration.

To be more specific, the problem I've come across recently is that it is possible for end users to create processes which pass validation, even though they shouldn't due to improperly configured gateways. When the instance terminates it is more than happy to tell you that you can't configure a gateway like that - but by that point it is too late as the instance is lost.

Sure, it is the responsibility of the designer to double check their work - but some of these business processes are maintained by multiple users from different departments so errors and mistakes need to be expected.

More importantly, as we pivot towards empowering our customers to build their business critical processes using this platform, we need to maintain their confidence in it's resilience even if they do make mistakes. At the end of the day, a single terminated process for a lengthy, multi-stage mortgage application can cost our customers more than just time and money - it can cause damage to their reputation.

As mentioned earlier, I'm just raising my hand to ensure it is noted if not covered elsewhere and to hear back from someone in the product management team to see if it is on the radar.

PS. I did do a search for topics related to this, but it seems the only solution for business critical terminations was to open a ticket and bring in Laserfiche support who run a script to restore lost instances. It sure would be nice if that script could be bundled in to the back end somewhere so that a forms server administrator could select and reinstate terminated instances (and there be an audit trail/log of the action).

Thanks again team! 

2 0
replied on November 4, 2019

I've discussed this same issue with Laserfiche support and can say that it is definitely on their radar. I think at this point they are trying to determine how best to address it in a way that covers multiple scenarios.

One idea that was floated in the discussion was the ability to move a process back to a previous step; in theory, this would allow you to move the instance back to a point before the error once the corrections have been made.

Again, this was all just a discussion of "possible" solutions so that doesn't mean we'll get this exact functionality, nor do I have any idea when it might happen, but it does tell me that they aware of the need and are actively working on possible solutions.

6 0
replied on November 4, 2019

Hi Jason,

Thanks for sharing your insights from your interactions with Laserfiche support. I'm glad it is on their radar.

Working through the problems myself I had certainly considered the biggest challenge facing the development team would be trying to implement a feature/tool designed to recover instances which could terminate for any number of reasons. That is certainly no easy task.

As you mentioned, I agree that a facility to 'undo' steps in a process would be a great place to start - so if that is on their radar, hopefully that is something we'll see soon.

Another hotly requested debugging tool was providing platform administrators with the ability to edit/correct process variables which may have been the cause of an error or suspended step in the first place. That would also certainly be of value..

The one thing I would like to reiterate for anyone on the team building this type of functionality is to remember to include some type of modification audit/history for both troubleshooting and support purposes.

3 0
replied on November 4, 2019

Kris, i agree 100%,

instead of the instance terminating, why not bounce it back to the last user action with a notification to a admin.

Or the ability to configure a outflow condition for exceptions to catch all instances that would otherwise terminate. 

Here's to hoping we see something soon. 

3 0
You are not allowed to follow up in this post.

Sign in to reply to this post.