UBE Performance suite - with a dash of cloud and AI

Understand your batch performance, immediately and over time.

Batch performance in JD Edwards is a strange one.  You only give it ANY attention when it’s diabolical…  If it’s reasonable then you leave it.  My clients start to get nervous about batch performance when they are getting close to the start and the finish of their batch windows.  Another classic example of batch performance getting attention is when scheduled jobs do not finish or there is a problem in the evening.

I like to be a little more proactive in this situation and have developed some insights with my team to allow you to quickly identify trends, oh – and then we’ve sprinkled a little bit of AI over the top to give you some amazing exception handling.  That’s right, AI in JD Edwards UBE processing – all will be revealed.

Firstly we need a mechanism of taking data out of the JD Edwards tables that are rich with UBE execution details, we upload them into google big query and then report over this data with some great dashboards.  We accelerate the value in this process by plugging each execution into AI and asking it whether this was a valid result – given the past results of that UBE. 

Firstly we have an agent that can run on any machine “on premise” that has internet access and access to your server map data sources.  It’d got some intelligence built in so that you can schedule it at a cadence that you like, and it’ll extract the relevant records and place them into some cloud storage [secured by keys and tokens and encryption and more]. 

I know a pretty graph is not normal in JDE (this can be hosted as cafe1 or e1page too) so that you see all of the relevant information at the source.

What this pretty graph can do is give you KEY metrics on all UBE processing, like rows processed, time taken and number of executions.  You have controls where you can slice and dice this interactively:

If you choose a particular environment (as above), user or date range, all reports and graphs are going to change.  You can look at particular queues or batch servers if you like

The example above shows the jobs for JDE and SCHEDULER and only the JPD920 environment – to narrow your focus.

We then provide a number of screens, depending on what you are after:

If you are looking for the history and trend line of a single job, you look at the job focus report:

We can see actual processing times, how many times run, who is running the jobs and how long the job is taking on a regular basis.  This is great trend information.  Also, we do not purge your cloud data – so you can do complete analysis on what jobs are running and who are running them – while keeping your ERP lean and mean.  We could even put your output in the cloud if you want – much cheaper storage!

I really like the graph above, this shows me ALL history of ALL jobs and how long they are taking on average and how many rows they are processing.  This is really valuable when looking for potential improvements.

See how many jobs are running at each hour of the day – knowing when the hot spots are for batch

You can look at your queues and find out what queues are quiet for the next processing opportunity.

You can get some great insights to solve performance problems, to know who is running what, and to keep your complete batch history.

Now for the AI

I’m a victim of technology, I want to put AI into everything – and this is a great use case.  We have the ability to look at things like return codes, rows processed, time of day and runtime and use AI to determine if the metrics are expected.  If the algorithms (that have been trained with ALL your historical data) think that there is an issue with any of those dimensions, they can raise an exception to you.  This is great for what I call “silent killers”.  If a batch job generally processes 40000 rows and processes 0 one night, it’ll still finish with status ‘D’ – yet AI is sure to determine that this is an exception and it’ll send you a message.  That is going to save time and money when fixing all the scheduled jobs that run without sales update having been run properly!  The nice thing about AI, is that is looks at the time of day and makes genuine decisions about the exceptions.

We run this as an end to end service, allowing clients to have access to all consoles and reporting.  We can also schedule any of the reports to be delivered at a cadence that suits.  Reach out if you want to know more about your batch processing!  There is a small monthly cost to the service.