Painpoints with IH-ETL - do you agree? any feedback you have?

[Suggy]

Issue 1:

We know that IH-ETL uses IRE, but its actually something like we are configuring/mapping the CI identifiers as part of the Attribute mapping step in CMDB integration studio.

ie say for computer, OOB the CI identifiers are like this

and this is what we map in the in studio. All the above identifiers are presented in the mapping step.

Suppose if you have 10 ETLs, you have to edit all of them in case you modify the CI identifiers. Suppose you had used Data source/transform map, you just call the IRE API and any changes done to CI identifiers will not be impacted, but in IHETL you need to update all those 10 ETLs

Issue 2- 

In Data source/transform map, you have the 'Choice action' to easily create/ignore/reject but the same is not avaialble in IH-ETL. You have to code it to do it.

Issue 3-

In Data source/Transform map, once the data is transformed you can see how many records got  inserted, updated, ignore etc.

TotalInsertsUpdatesProcessedIgnoredSkippedErrors  

The same does not work for IH-ETL.

[-O-]

Actually 3. is gray area: the official paradigm - build payload using transform map, call IRE, cancel row transformation, leads to the same situation: no immediate true information about the result of a transformation of a row is available.

So using the official paradigm, to update the CMDB from a data source using transform map and IRE is no better - everything is shown as ignored.

I said gray area, because one could come up with a scheme where each data source is handled by two transform maps:

- one that does the IRE part, which has just one scripted field map: sys_id - which will be used for coalesce

-another that functions as the payload builder part - which would actually be disabled/not-active.

Thus in the active transform map one could run identification and could set transformation result to: update or error.

It would never be insert, cause the insert would be done by IRE during coalesce execution.

There may be other ways to to obtain "real" transform results during transformation and actually would be happy if other shared their solution.

[Suggy]

Anyone else

[Suggy]

Anyone else has any inputs?

[Suggy]

No one?