add stresstest cluster command

[butonic]

continuously do one of the following db queries

• update and get - to check if a read write split is working correctly
  • oc_preferences is fine for this
  • generate unique id so multiple processes write different entries
• update and get in transaction - to check if read committed is working correctly
  • oc_preferences is fine for this
  • user same tupel for all processes
  • wrap in transaction
• insertifnotexists as in the cache