Any info on spikes in SQL DB capacity consumption?

[arpost]

Greetings, all. We have an F64 capacity and have been testing out Fabric SQL DBs with some heavier data ingest processes and have been noticing some odd results in terms of massive spikes in resource consumption. These are shown as interactive usage on the Capacity Metrics report despite no operations being actively performed in the DB by any user.

Here's an example from a recent pipeline run. I was bulk copying files (file sizes range from 45kb to a few that are 140MB+) into SQL DB stage tables. Nothing fancy as the pipeline was only doing a Copy activity from .txt files to stage tables in the SQL DB (no data typing on fields being applied).

I cancelled the job at 11:43 AM, but Capacity Metrics show massive spikes on the same DB 10 minutes later (and continued to register them until at/after 12 PM). These operations ate up majority of our avaiable capacity despite no one accessing/using that DB.

Are there operations that Fabric SQL DBs run in the background that account for this? Are there ways to control/limit these as it seems like loads of this nature should be able to be handled.

[rohitdh]

Hi @arpost, I am from the Fabric SQL team, could you share the item id of your SQL DB so I can investigate the telemetry around this time period?