Nicholas Blumhardt

Kicking off an experiment in https://github.com/nblumhardt/serilog-enrichers-demystify

The _Serilog.Enrichers.Demystify_ package mutates the original exception, so it won't avoid this issue (the stack trace etc. will still just be reset by the re-throw). It seems like the best...

> Mimicking original exception with the subclassed one (ExceptionInfo) still could create behavior change in downstream sinks when is operator won't work as before along with the GetType() and other...

Thanks, Dani 👍

This is very similar to datalust/seq-tickets#496, which addresses a similar issue deeper in the ingestion process by converting any floating point too big to fit Seq's underlying `decimal` numeric representation...

This needs deeper work than we can squeeze into the final few days of 2021.3, so moving this out of the milestone for scheduling in our next triage.

There are some substantial hurdles on the server-side that will prevent us from addressing it directly in Seq, but here in _Serilog.Sinks.Seq_, (and we think in the other client libraries,)...

Thanks for the heads-up. That's not our code - you'll need to raise this on `dotnet/runtime`. If you need any extra info though, please loop me in :-)

CC @snakefoot, who may be interested in this from the NLog perspective.

:+1: great idea.