Hi Elementary team, what is idea behind backfill-days limited to 1 as smallest? Say I run volume-anomaly on table and have time bucket period of 4 hours and dbt test job is running multiple times per day, having 1 viable fail at 4am I will continue to see test(and job) fail till the end of day(or maybe even 24h) having data flow recovered.

Hi @Dmitry Darkin, This is currently a gap we have, but there's another workaround you can use here which is the suppression_interval param. This allows you not to get multiple alerts on the same issue for the number of hours of your choice. read more here

got it, thank you. My situation includes failed dbt job alerting too. For now I put only drop direction for anything less than day.

@Dmitry Darkin do you mean this spike that was marked as an issue was also a false alarm?

no, it was viable. I just do not see spike as bad anomaly, whereas drop is mostly bad. So do not want job fail on spikes.

ah ok got it. yeah in Volume anomalies it's very common to alert on drops only 🙂 In any case, I hope the suppression interval will help you with the issue of repeating alerts