I understand the need to be able to check at regular interval whether the entropy source is still correctly functionning but rngd wakes up hard drives from ther sleep state every time it has to dump stats to the syslog.
Could we have the same parameter as rngtest:
-t n, --timedstats=n (default: 0)
Dump statistics every n secods, if n is not zero.
[quoted from man rngtest)
This way we could set a larger interval if needed.