Wednesday, 29 April 2020

load testing orchestration - the cheap way

I have a client with an orch that is sometimes failing, the problem is that the previous order number is not always returning a value, so I'm trying to determine what is going on.  The way I'm going to do this is simulate numerous orchestration calls.

I'm sure there are heaps of ways of getting this done, but I'm going to demo something really simple!

One script with the following contents.

It'll run 10 times, you can just change the number 10 to what ever you want (in the while control)

You can do your own timing changes, and you can also text return codes - just google curl!

This is using basic auth, and yes - I've changed it so it's not going to work for you, haha.

My orchestration is called orch_AddS1WarrantyOrder2

The default payload for the orchestration is in the variable $data

I need to change the PO every time, hence modifying the variable with the PID of the script and also using an index.


Super simple!  paste the contents, chmod it and run it!

data='{ "inputs" : [ {    "name" : "orch_inputShipTo",    "value" : "379015"  }, {    "name" : "orch_inputMCU",    "value" : "MDC"  }, {    "name" : "orch_inputCustomerPO",    "value" : "49354384MK7"  }, {    "name" : "orch_inputQuantityOrdered",    "value" : "1"  }, {    "name" : "orch_inputItemNumber",    "value" : "90110263"  }, {    "name" : "orch_inputSubLedgerType",    "value" : "I"  }, {    "name" : "orch_inputSubLedger",    "value" : "00010155"  }  ]}'authString='Authorization: Basic SkR9889781pbiQ='
i=0
while [ $i -lt 10 ]  do 

  data2=`echo $data | sed "s/49354384MK7/$$384MK7${i}/g"` 

  echo $data2     

  curl -v --request POST  \          --url 'http://10.10.20.24:8081/jderest/orchestrator/orch_AddS1WarrantyOrder2' \          --header 'Accept: */*' \          --header "${authString}" \          --header 'Cache-Control: no-cache' \          --header 'Connection: keep-alive' \          --header 'Content-Type: application/json' \          --header 'accept-encoding: gzip, deflate' \          --header 'cache-control: no-cache' \          --data "${data2}"  i=$(($i+1))
done



Thursday, 23 April 2020

managing a remote workforce

I'm hosting this webinar.

If you are an existing ERPInsights customer or are interested in how to gain insights about waht your users are doing in JDE and how things have changed since they are working from home - please join in.  Even if you want to heckle from the back row - that is cool with me.

I promise a zero contact webinar!




Wednesday, 15 April 2020

JD Edwards Technical debt reduction - Favourite flavour is Vanilla



1.1       Introduction

Fusion5 has developed specific offerings to assist clients to firstly, recognise their technical debt reduction opportunities and secondly, actively reduce them.  The reduction of technical debt allows clients to be more agile and stay code current at a reduced cost.

This white paper explains what technical debt is, strategies to reduce it – and of course how Fusion5 can help.

Getting your JDE code back to vanilla is exactly what we’re aiming for.

This is a project that you can take on anytime and dedicate any number of resources to.  A project like this is going to save you money - so perfect to take on at a time like now (pandemic).

1.2       The true costs of modifications

Not only that, we recommend that our clients make it their favourite flavour too!

What am I talking about?  Keeping your JDE code as vanilla as possible.  Only change standard code if it’s going to make a strategic difference to your business.

The cost of a modification grows and grows.  If you look at your modifications, especially if you are modifying core objects – retrofit is going to continue to cost you money going forward.  This is more apparent when you modify an object that changes often.

Let’s look at those with a free dashboard we have created:





For example, P42101 has been updated 10 times since 2016.  If you have made modifications to this object, then this is going to cost you 10 x the retrofit costs.  Or at least a factor of the original costs every time you try to be code current (which is the new normal).

Hours
Cost
Original Modifications
15
$3,000.00
retrofit after JN12140
7
$1,400.00
retrofit after JN15493
6
$1,200.00
retrofit after JN16709
5
$1,000.00
retrofit after UN3
4
$800.00
retrofit after UN4
5
$1,000.00
TOTAL
42
$8,400.00

My very simple example above shows that the original cost of the modification might have been 3K, but the ongoing maintenance of this has cost an additional 5.4K – nearly double the original modification cost.  These are big round numbers for illustrative purposes – but you get it.

This is also an inhibitor to change (X factor or something hard to measure), instead of taking all of the latest changes – we tend to hold off because of the impact to the business.

The lesson is simple, we need to reduce our modifications, period.

1.1       Modification categorisation:

We like to categorise modifications in 4 distinct categories.

Category
Description
Modify standard object
Any changes to a standard object will continue to require retrofit.  These are generally identifiable using the C flag in F9861. A more advanced method is looking into OMW history to ensure that the modification is real.  The Super advanced method is comparing code.
Copy standard object
If you have created a custom object – R5542565 to do the job of a standard object – this can cost you.  The caveat is, will you ever want the new base functionality in your copied object?  You need to be smart with cat codes to make this decision.  If objects are clearly marked as “bespoke – don’t care about upgrade / update” – this simple annotation is going to save a lot of money.  In general, if you copy a standard object, this is going to cost money moving forward – as you’ll want to copy and retrofit again when the base object changes.  These are generally identifiable by name (hopefully a copy of the original), name and system code conflict (if the original system code was used) and finally looking into OMW logging history.
Bespoke with standard hooks
It’s yours but it is called by or calls standard objects…  It might read standard tables and rely on standard inputs.  If these change in an ESU – then there is going to be a roll-on effect.  These are identified by name and system code (hopefully) and also cross reference for seeing whether there are standard hooks in and out.
Bespoke
It’s yours and it does not matter what happens around it – bulletproof!  These are identified by name and system code (hopefully) and also cross reference for seeing whether there are standard hooks in and out.  There should be no references to standard objects.

Once you have categorised your modifications like this, you can begin to determine which ones need to be kept, and which ones can be converted to a debt free mod.

1.2       What is a debt free modification?

A debt free modification is one that is stored as configuration (not code).  In general, these are created “on the glass” and survive updates and upgrades.  There have always been ways of creating configuration-based modifications.  Grid formats for example were a classic way of customising your environment without making development changes.  This type of “citizen developer” customisation has matured into the UDO User Defined Objects that we have today.

We all need to be experts in UDOs to really take advantage of them.  It helps when developers know when not to write code, and when functional people know when to engage developers.  We recommend training your team to ensure that they know about UDOs and have the ability to implement them.

Also, start your learning journey here:  learnJDE.com

You can get a jump start on learning by watching some quick videos and then trying things out for yourself.



In terms of converting your modifications to configuration, unfortunately there are no super easy ways of doing this analysis – quite often it’s a lot of hard work.  There are a number of techniques to assist you though, which we are going to cover.

1.1       Finding debt free modification opportunities

Firstly, your development team should be able to assist.  Start with your changed objects and work backwards.  Really, you only need to worry about APPL, UBE, NER and BSFNs – as they are going to occupy 90% of your retrofit time.

High level categorisation (as opposed to knowing or inspecting every object) is super beneficial to understand the scope of a project to either retrofit or convert to configuration.

There is some help from Oracle.

Tools Release 9.1 Update 2 (TR 9.1.2) introduced two new utilities in an aim to simplify the upgrade and patching process by reporting on “what’s changed” in your environment. The Customization Object Analyzer provides summary and detailed level reports on customisations made to JD Edwards EnterpriseOne objects such as applications, UBEs, business functions and business views. The Application Configuration Analyzer compares two environments and reports on the differences in UDCs, Processing Options, and Versions.

The Customization Object Analyzer utility can be used to list EnterpriseOne objects that have been customized. This information is then used to determine the impact and level of effort needed to retro-fit the customizations into a newer application release of JD Edwards EnterpriseOne software.

The Application Configuration Analyzer is a set of reports which identify differences for processing options, data dictionary (DD) items, and user defined codes (UDCs) between releases or environments.

Tools Release 9.2.3.3 introduced the JD Edwards EnterpriseOne Decustomizer Analysis Report which enables you to identify the UI customisations that can be replaced using the personalisation and extensibility framework.

This document lists the concepts, requirements and available information on Customization Object Analyzer, Application Configuration Analyzer and the Decustomizer Analysis Report.

Some additional observations of debt reduction are below:

Sample modification
Technical debt reduction technique
Form exits and row exit customisation
Form extensions and form personalisation.
New button or exit
You must use the new orchestration functionality to associate a button to an orchestration.
Hiding fields / simplification
Security for form personalisation.  Grids.
Pre-populating fields
Advanced Queries
Showing additional information from other JDE screens
Cafe1
Bringing external content to a page
Cafe1, orchestrations and form personalisation.

Any modification that you convert to configuration is going to pay you back double.  You’ll be able to take change more rapidly and you won’t need to worry so much about mobilizing a development team.  You must understand all configuration options available to you and cross reference them with your modifications.

1.2       Often forgotten customisation workbench (P96UCUST)

When you are into the mechanics of technical debt reduction, this program is going to be your friend.


This is an often-forgotten gem when it comes to retrofit, and it does have links to the latest functionality, in that you can get to the “Object Customizations Data Load Report” and xref and other functionality quicker.   You’ll notice a theme here; this is exactly how we’ve categorised our modifications earlier.



  1. Refresh All - calls BSFN to refresh the F96UCUST data based on the merge modification flag in the F9861 table.
  2. Load COA Data - launches batch versions for the Report R9840DL - Object Customizations Data Load Report.
  3. Launch XRef - will launch the entry point form of the cross reference application in order that the end user may gather more information about the currently checked in object.
  4. Launch OMW - will Launch OMW (P98220) where more actions can be taken on the object and project as needed.

At Fusion5, this is ground 0 for our retrofit efforts.

1.1       How can Fusion5 help reduce technical debt?

There are two things we do:

1.       Reporting - to let you know what is possible.
2.       Converting – remove technical debt and standardise documentation.

1.1.1       Reporting on technical debt

We’ve created our turnkey solution that enables the best results for our clients to visualise their technical debt.  We have developed custom queries and dashboards over the last 15 years which allow us to understand modifications intimately and then make specific recommendations to remove them.

We extract data from a series of tables in JD Edwards and load this into the cloud.  We run custom reports and also load these into the cloud for easy and interactive reporting.  We load up cross reference too, so the results are complete.


At the end of the exercise you’ll have a complete macro and micro view of every object in your system and how you can take these back to vanilla (where possible).

We can even extend this service to work with you on each modification to understand whether this has been included in an ESU or ASU, and retirement of the change is easy.


You will feel empowered to see how your technical debt can be reduced and therefore make you more agile in your capability to react to change and stay code current.

Just reach out and ask about our technical debt reduction reporting packages.


1.1.1       Converting Technical debt

These projects can run all of the time in the background.  A technical debt reduction project can run with a single developer, using the reporting from step 1 and converting code to configuration.  At the end of the day, everything should continue to work as it did.  The technical and functional specs are easy, and the test plans are easy – things just need to work – as they did before. 

You can work through objects methodically.  You could select a system code at a time, which would allow for better ESU application for particular high traffic modules.  The only goal of the exercise is going through changed objects and taking out the modifications where possible.  And if that’s not possible, using category codes and documentation to ensure that this IS easy next time it occurs.

Fusion5 will focus on these affected objects and ensure that the modifications are as debt free as possible.

Debt reduction programs can be done by object, or can actually run as an agile project, which would be based upon how many modifications could be retired in a 1 week sprint, and continue.

The added advantage of this project is that it will save you money as it runs – it will save more than it costs – some nice ROI.


1.2       Conclusion

Fusion5 has invested heavily in code and knowledge around reducing their clients’ technical debt.  We’ve designed programs to make it easy to identify and educate our clients on technical debt reduction. 

Our capability in this space is sure to leapfrog your ability to adapt to change, to change quicker and ensure that change does not cost too much money in JDE.  Our goal is to enable our clients to be on the latest code all of the time, and take advantage of what Oracle is putting into their product.

If you want to start the code conversion journey, please get in contact so that we can assist you.

You can download this as a PDF here:  



Monitor JDE usage patterns and changes during COVID-19

I'm hosting a webinar on using ERP insights to ensure that users are getting a consistent experience with JD Edwards - whether they are working from home or not.
Register and listen in.  This is really good for both JDE customers and JDE partners, as we have a channel model for you to assist your customers interpret this invaluable data.

Learn how we use this dashboard and more:



Wednesday, 1 April 2020

Some of my favourite screen shots.

A really quick snapshot of the kinds of dashboards we are providing our ERP Insights clients from around the world.

Our clients are

  • Doing SOE (Standard Operating Environment) and compliance checking using this information, especially with everyone working from home.
  • Validating that JDE is being used effectively and continuing to perform for remote workers
  • Getting a good idea of what are critical business functions within their organisation during times of duress
  • Able to compare long term and short term usage differences for important trend / macro changes in performance and behaviour.
I'm never writing a PDF report again, I'm going to use dashboards for everything.




Our dashboard - designed to assist you identify the programs that need testing per system code.  You are able to select the system code and see all of the programs that need testing.  You can view this data for any time period and any environment.

You are able to see the users per module in the time period that you choose.  This allows you to see how many users have access programs in certain modules.  Note that this can easily be broken down to user level data which can include version and form information for JD Edwards.

Here we can see a really simple summary of how many apps are used per module and how many users (unique JD Edwards user id's) have accessed that module in the time period specified.

This is a month on month comparison of JD Edwards usage from around the world.  You can see an exact daily compare of page loads and choose the module or country that you are interested in - on the glass.  This allows you to see which modules are changing their use and also which regions.

This is a list of batch jobs run in the last week, and how that compares with the previous week.  The comparison data points include rows processed, time taken and number of times run.  This is an easy way of comparing your batch activity week on week and ensure that things are consistent.

A complete UBE summary for 3 months.  How many jobs are launched daily and what is the average queue time, run time and rows processed for every single batch job.  

Queue breakdown by UBE, find which queues have the most wait time.

Which programs are the busiest from around the world.

Extending JDE to generative AI