Hi Everyone, We have configured certain elementary freshness anomalies tests for our DBT project. Below is the configuration for one of the tables. For an entire week, before the ingestion process was eventually fixed yesterday, we failed to load this table using the dbt model. The question that I have is, the freshness test for this table on the dashboard does not show any failure. I am assuming the configuration needs to be tweaked on few more parameters. Could someone please shed light on this. Attaching the logs from the dashboard graph as well. Could you please direct to me any documentation on how the freshness value, and expected range are actually calculated

I believe the elementary freshness test does not account for dates that are missing entirely

Hey @DE-2, how often normally this table is updated ?

Hi Daniel,

And can you share a photo of the full graph ?

Do you need more Data Points?

Yes, to understand you case better. Basically the freshness value is the amount of time we didn't see an update. And as you can see, there is an incline in the last points in the graph, which indicates some delay in the update

Yes. But the idea is to raise an anomaly even if a single day has been missed in loading the table.

@DE-2 you can try to lower the sensitivity of the test (lowering this parameter will make the test more sensitive)

Right. I was thinking the same. I would need to trial and error with reduced anomaly sensitivities to raise anomalies more frequently. Thanks @Daniel Pollak I will let you how it works. Any other parameters I should be looking at tweaking?

Not that I can think of. Sounds good 🙂