benchmark single list vs iterator for active fps (DNM)

[KonradStaniec]

Compare storing active finality provider as list with iterator vs as list in db;

Results for 80fps (on my machine):

80 FInality Providers
Get all finality providers
Old: 20243087 ns/op
New: 19692 ns/op // 3 orders of magnitude faster

Get one finality provider
Old: 1072 ns/op
New: 7441 ns/op // 7 times slower

I am still not sure whether all data is flushed from memory to the disk, but in setup I do:

• commit which should flush the data from memory to the goleveldb
• write pretty large array to goleveldb which should flush internal lsm memtable to the disk