3807720 - Unexpected spikes in data backup size without corresponding data growth

• The complete data backup size was observed increasing but there isn't large table size change. 
• SQL "HANA_Backups_SizeEstimation" enclosed in note 1969700 shows that the estimated complete data backup sizes is indeed larger without table growth.
• Following messages can be observed in indexserver trace:

   KernelSentinel   KernelSentinelJob.cpp(01691) : There are too many un-collected versions on table '<schema name>.<table name>' (TABLE_OID = ***, NUM_VERSIONS = ***, MIN_READ_TIMESTAMP = *****).
  The cursor possibly blocks the garbage collection:  
   CONNECTION_ID = ***, CONNECTION_STATUS = IDLE, HOST = vhzjehp4db01 : 30240, IDLE_TIME = **** sec,
   STATEMENT_STATUS = SUSPENDED, STATEMENT_LIFE_TIME = ***** sec, ISOLATION_LEVEL = READ COMMITTED.       

• Following HANA alert is raised:
  

  Priority:
  4 (High)
  
  Alert ID:
  42
  
  Alert Name:
  Long running/idling cursors

• SAP HANA, platform edition

hana backup size spike, complete data backup large, cleanup file size, long running cursor, idling cursor, suspended statement, garbage collection blocked, idle_cursor_lifetime, mvcc history, hana_garbagecollection_overview, hana_backups_sizeestimation, jdbc client, transactional lobs, backup growth without data growth, cursor timeout , KBA , HAN-DB-BAC , SAP HANA Backup & Recovery , HAN-DB , SAP HANA Database , HAN-DB-PER , SAP HANA Database Persistence , How To