Great response, thank you!
I’d love to go the Kapacitor route instead - much easier to edit and maintain the scripts. Unfortunately, after many, many hours of experimentation and discovering of bugs, I’ve come to understand that Kapacitor is no good for updating historical data. Say I improve my query and discover the interesting stuff is <= -49, not <= -50. With CQs, I can run the query without the CQ wrapper and it’ll run on all my data, overwriting historical entries. With Kapacitor scripts, attempting to record/replay is a minefield of unexpected behaviour (see, for example, Kapacitor record query does not return same points as line protocol query · Issue #1294 · influxdata/kapacitor · GitHub).
