Foivos Zakkak

> @zakkak I think we should improve how the metrics are collected so that we are able to provide a diff of what's new. Otherwise it's very hard to take...

FWIW in this case it seems like the 1% threshold for JNI methods is just a bit tight...

Looking at the provided build output one can already see some things:: 1. The time spent in the analysis is way longer in 23.1 than in 23.0 and that's reflected...

> FWIW, I know that at some point a Kubernetes Client update caused some problems as the amount of code was massive. Correct, that was https://github.com/quarkusio/quarkus/issues/37142

@vsevel looking at the 3.5 vs 3.7 build outputs (I am assuming you are using Mandrel 23.0 in 3.5 and Mandrel 23.1 in 3.7) I see that both the code...

@vsevel I forgot to mention that in order to reduce the moving parts to the minimum it would be better to build only using Mandrel 23.0 for the time being.

@bredlej judging by the error message: ``` Error: fork/exec /var/task/bootstrap: exec format error ``` I suspect that what's happening is that you are building a native image targeting your Mac's...

> Do you have perhaps any ideas what to try out? Unfortunately I am not familiar with AWS lambdas and the referenced scripts, so someone else will need to step...

@sberyozkin IIRC the same error is thrown in the `bouncycastle-fips-jsse` (last time I checked) integration test (I didn't update https://github.com/quarkusio/quarkus/issues/14139 though...)

> Runtime reinitialize `org.bouncycastle.jcajce.provider.BouncyCastleFipsProvider$PooledSecuredRandomProvider` That only works for static fields that are initialized at class initialization, it doesn't help for object instantiation. > Reset the field Not sure why you...