App locked inside the core emulation thread

[sergystepanov]

There is a situation where a core can become stuck on the same valid or invalid operation, effectively locking the worker in such a way that it remains active for liveness checks and connections but remains perpetually stuck. This state persists indefinitely, or until the log size (if it is writing logs for these operations) grows large enough to crash the system.

A possible solution might involve adding some verification to ensure that the cores are still producing normal emulation data. As for the specific conditions — who knows. :/