Custom problem intergration

[henk_stobbe]

Hello,

When using the custom problem integration, I notice that by design, {Tags} and {ImpactedEntities} are not JSON but very inconsistant CSV is used instead. This means that any JSON parser stops at this point and you have to parse the rest on your own.

You get keys, keys with values and keys starting with [] all seperated by ",".

My intention would be to convert from CSV to records, and then splitup by ":" finding my key and returning the value that it contains.

Nicer would be that dynatrace would send all JSON, just add a dummy value to keys without value and do not use the []. characters, and ofcourse put {} around the CSV.

Does anybody have any thoughts about this?

KR Henk

[Julius_Loman · Accepted solution]

With the tags - you are right, it's not a structure, but a string of tags separated by a comma.
However, the ImpactedEntities is an array of JSON objects thus you don't have do any special handling.

With tags - They are accessible in a structured way in the {ProblemDetailsJSON} placeholder in "tagsOfAffectedEntities" attribute.

[henk_stobbe]

Hi Július,

Great answer thanks, now I see the difference between ImpactedEntity and ImpactedEntities

And I norice the "new" in my 162 instance(-;

KR Henk

[Julius_Loman]

These "New" placeholders are there for quite a long time - from Dynatrace 1.143.

[henk_stobbe]

Then this is strange (-;

[henk_stobbe]

example picture

[Julius_Loman]

Yes, it shows them as new - but there are "new" for about 10 months. 10 months is a very long time for Dynatrace 🙂

See more in this blog post.