Separate stream+window for each tickscript or combined tickscript with single re-used stream+window

kapacitor
#1

What is more efficient :

  • A single tickscript with single stream() and window() nodes where all (potentially logically separate) alerts re-use those nodes
  • Or multiple tickscripts having exact same stream() and window() definitions, and its own alert() logic

In the first case I can guess there might be a performance benefit, but to which extent? Aren’t streams+windows, having the same definition, got re-used internally even if they are declared in different tickscripts?
Where the first way, apparently, is not that convenient for alerts organization and management, and statistics consumption.