Retry insertion errors on ingester

[MarceloRobert]

The ingester is one of the most important tool, and with upcoming changes some unexpected errors could happen. Even still, we are currently simply not dealing with insertion errors (cf kcidbng_ingester L384) and this should be addressed.

[MarceloRobert]

On hold while we check if Denys will be the one to work on the ingester

[tales-aparecida]

With more and more constraints,

Why were we thinking of adding constraints? Can we find/create a ticket with that discussion?

[MarceloRobert]

@tales-aparecida There is a newer issue for foreign key constraints on #1514

But this issue here could be more general, outside of constraints, since currently we don't do anything if the ingester insertion fails on any kind of exception. What do you think?