Unified alerting: range(start: v.timeRangeStart, stop: v.timeRangeStop) possible bug

[kortenbach]

When defining an alert query in InfluxDB Flux script I use:
range(start: v.timeRangeStart, stop: v.timeRangeStop).
When I test the alert in the editor it works as expected.

However: The resulting alert doesn’t work in “normal” mode. (It always evaluates to health:“nodata”).

Example: “range(start: v.timeRangeStart, stop: v.timeRangeStop)” with a selected time-period of “now-3h to now” results in health “nodata” whereas range(start: -3h, stop: now()) results in a working alert (however: In the alert editor the query result will display an error now (“Data outside time-range”) with a button “zoom to data” that doesn’t work.

Again: While editing the alert using “range(start: v.timeRangeStart, stop: v.timeRangeStop)”.everthing works fine. The selected time-range is reflected in the query result graph and the alert test works.

I suspect that v.timerangeStart -Stop are not working properly when not in the alert editor.

[jamesfd70]

Did you ever find the answer? I’m having the same issue…
My problem/solution was a selfinflicted typo !

[kortenbach]

Hi James,

Long ago. I don’t think I found the solution. Started using Grafana alerts instead…
Good luck!

[grant2]

Hi @kortenbach and @jamesfd70

Grafana alerting has been updated a few times since then. In the examples that I created in this tutorial, I never had an issue with the time selector (and I tested pretty rigorously!). I basically always wrote my queries using

|> range(start: v.timeRangeStart, stop: v.timeRangeStop)

Example: