Scenario Design / Management - Best Practices?

[KatherineLa]

We're new to Fusion, implemented earlier this year. We made immediate progress in automating a number of repetitive steps/errors that happen when new projects are created by our PM team. They've started to pick up on how it works, and we're getting requests for customizations for particular client/project needs as well.

For some clients, we do a high volume of repetitive work, so customizations absolutely make sense. What I'm looking to understand is if there are any best practices for structuring the scenario design so we can track what happens, when and where before we make a mess.

Ex - today I have a 'New Project Setup' scenario that executes a collection of checks/updates that we want run on every single project. Then we built one for a specific high-volume client that does another lengthy collection of steps but also triggered by a new project. I see a world coming where I have a lot of these, so I'm looking for ideas on corralling them all. Any watch-outs?

@Mike Haleas‚ Tagging you to watch this one too.

[stuartcairns]

Hi Katherine. I think simplicity is key in design. For me, if it's going to take some thinking then I built my scenario 'longhand' and then when each stage is defo working, I go back and see if I can simplify and remove any modules perhaps by gathering data in single steps and making fewer calls to sources. Once it's as simple and efficient as I can make it, I then look for places where things may go wrong, for example, a missing bit of data. Immediately after that I put in a router to send an email if the 'failure' occurs. That way I know if anything goes wrong potentially before the user. You could use a similar thing (without the router) to write a record into Excel, GSheet etc to document the run if you like. You can save a comment, maybe the data or part of it. Essentially, you can make your scenario's self documenting. Nice!

[NathanHa1]

We've put together a couple of monitoring scenarios that track potential incorrect and high-impact/high-value errors and notify those affected and a Fusion admin when it happens. It's hard to know where issues will arise, but if you start with a couple worst case scenarios and monitor them, over time you may be able to identify the trouble areas and build specifically to those issues.

Another thing we have done is implemented custom fields/forms just for tracking purposes for certain processes. By writing back an id or label of some sort, it could make it easier to identify where the issue came from.

Also, the advanced search options on the history tab can do a full text search making it easier to identify the executions in question when troubleshooting.