Thursday 19 March 2020

UBE analysis - long term trend

Something that has given me grief for a long term is understanding the impact of a tools release, package deploy or any technical change on performance.

Another thing I've struggled with is keeping enough batch history to understand my performance trending.

Well - have I got a post for you.

We've been working hard on taking JD Edwards data offline into the cloud - and providing generic reporting over the top of this data.  We are starting with WSJ, but are going to move out to security history and perhaps work centre messages.  What we are going to be able to provide is a seamless window to that data quickly, easily and with more insights.

You'll see below how we use a secure agent to transfer JDE data to bigquery (encrypted at rest and in transit) and provide our clients with a complete reporting suite into their batch history.

So back to my first two pain points - I can keep all the history I want - because it's in the cloud.  And secondly, I have reports like below that allow me to drill down to those "all important" historical cubes to see what the performance of ANY UBE has been over time and reconcile changes quickly.

A demo video can be found here


And here are a couple of screen shots. 

This shows that I initially drilled down on system code to see all of the 31 UBEs, I then chose a single UBE and then a single version.

I wanted to look at the daily average since Nov1 - and there we have it.  A graphical summary.  You can see that performance is pretty consistent.  The below graph represents the rows processed and time taken, which have a symbiotic relationship.



 IF you want to run this as a 1 off with all your UBE history, we can arrange that very quickly.  If you'd like us to carve off your data and keep JDE quick, we can assist you with that too!


No comments: