We're seeing a problem in production which looks to be stuck in 'receiveLookupOrCreateResultByProps' or 'receiveRegisterResultByProps' causing everything to be stash.
I'm looking to fork glokka to add a stash size monitor and formally name all actors so we can monitor with Kamon, also a reset which will flush stash and become receiveGlokkaEvents
In the meanwhile, Is there a scenario you can think of where Found/NotFound/Conflict/Registered are not received?
Our Registed actors has very slow preStart (~60s) could a timeout be kicking in and not handled?