How to generate Pentaho documentation?

Question

Our organization has used Pentaho for data integration purposes for over a decade. Due to organizational changes earlier this year, our team should take responsibility for data integration going forward.

The problem is, our team has nearly zero experience with Pentaho and use Frends for data integration. We want to move everything out of Pentaho to Frends. But alas, Pentaho seems to be a fairly challenging customer.

Problem 1: There's a lot and I mean A LOT of shit in Pentaho. The depth of some jobs is quite staggering, at least by our standards.

Problem 2: There is zero external documentation.

Problem 3: There is almost zero documentation within the jobs.

Problem 4: Pentaho is slow. Switching between open tabs easily takes 10 seconds. Opening a new tab 20 seconds, closing a tab 10 seconds.

Problem 5: A job within a job may be named as ABC, but when you open ABC, it is actually named as DEF. Combine this with Problem 4 and trying to keep up where you are, and where you were is hard.

At first I tried to document relevant jobs into our wiki using a tabular structure. But it quickly became a mess that was nearly impossible to navigate. Considering the outcome, spending any more time to manually document Pentaho seems stupid.

Is there a way to generate (hopefully readable) documentation automatically? I googled for answers and found this page. But considering the zero experience with Pentaho, I do not know if that page describes what I am looking for, and the instructions are written in rather broad strokes I have no idea how to apply in practice.

Thank you in advance.

Answer 1

Re your problem 4: there's no reason for Pentaho to take that long opening or switching tabs. Unless you're running it on particularly slow VMs with little memory, it shouldn't take that long. It's not the quickest gui out there, but it should be fast enough to allow working.

The only documentation tool I know of is the one you mention on your post. But as far as I know it's quite old and I don't know how well it'll work for your version. Give it a try, and see how it goes.

Other than that, if there's no doco available, you have litle recourse other than going through the various jobs and sub-jobs, transformations and sub-transformations and reverse engineer them.