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.

[Jerome22]

I have the same issue.

just opening the performance dashboard of an SQL Database causes the spikes

I strarted from 0 CUs consumed to 1600 just by navigating the performance dashboard.

crazy.

[MrSkeie]

Hi! We are facing the same issue, SQL Database has constant capacity utilization, around 12% of our F16. For the last 24 hours, we have had zero CRUD-operations. PowerBI sync semantic model on SQL analytics endpoint is turned off and there is absolutely no activity. Still, the CU is constant at all timepoints for the last 48 hours.

[rohitdh]

Hi, sorry to hear about the CU usage problem you are having, we will need more details on your scenario and resources to investigate, could you please open a support request: how to create a fabric support request

Once you create the support request, please share it offline and I can send it directly to the team.

[Henson]

Similar issue here. In our Capacity Metrics app though it really looks like it's showing as "billible" and we got the "You're at 100% and will be throttled e-mail" right after it spiked. 

[rohitdh]

Hi @Henson, could you please open a support request: https://learn.microsoft.com/en-us/power-bi/support/create-support-ticket and share the ticket offline?

[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?