#1893 fix ensuring the consistency when doing signal enrichment

[thjaeckle]

Resolves: #1893

[kyberpunk]

Hi @thjaeckle you are very fast, looks perfect. Just one question: how exactly we can enforce the new behavior? Should we somehow setup ditto headers? Thank you.

[thjaeckle]

@kyberpunk I would make this the default behavior - we want to ensure consistency here
There is also an optimisation in place, which does not apply the rather cost-expensive "history lookup" - when there is no newer modification of the thing, the "history lookup" is skipped and the in-memory representation is used for retrieving "extraFields".

That should for 95% of cases optimise and only for high-frequency cases add some additional cost.

[thjaeckle]

@alstanchev @kalinkostashki could you please have a look?
Would want to release a 3.5.3 soon

Could you please run system-tests for this PR as part of a review? That would be super

[kalinkostashki]

@alstanchev @kalinkostashki could you please have a look? Would want to release a 3.5.3 soon

Could you please run system-tests for this PR as part of a review? That would be super

hey @thjaeckle,
sure will do :)

[kalinkostashki]

@thjaeckle, system tests are green
and thank you for re-adding the retrievePartialThing method

[kalinkostashki]

LGTM 👍