The downsampler currently accepts a query period and then runs aggregates over that time-range.
However, a lot of my Flux scripts query a long period and then window it - the idea being that if a run fails, the next will catch things up.
It'd be good to support the same - specifying a window period in YAML
Activity
04-Feb-23 20:47
assigned to @btasker
04-Feb-23 20:48
One of the challenges with this will be calculating the timestamp to use with each window.
Currently we figure it out from the task start time and the query period, that won't work with a window
05-Feb-23 09:52
So actually, adding this will remove the need to calculate the timestamp -
_start
and_stop
get aligned to the window.05-Feb-23 10:02
mentioned in commit d79295eea23c4b31a59527257fe393895f906cce
Message
Don't custom calculate time windows, have Flux do it for us utilities/python_influxdb_downsample#5
This switches to using
window()
to calculate windows, allowing much longer time periods to be queried and broken up into smaller windows05-Feb-23 10:02
mentioned in commit cca6226f8982c26c0d81100cf576c776bc3bc5ec
Message
Add support for
window()
utilities/python_influxdb_downsample#5If
window
isn't included in the YAML, it'll default to be the same asperiod
05-Feb-23 10:04
mentioned in commit c19e4bf5914388b3f25776353346e274cb45305d
Message
docs: update README utilities/python_influxdb_downsample#5