We have Azure SQL on S3 level.

For about the last 3 hours, we've been hitting 90%+ DTU, with occasional dips down to <10%. Here's the query I'm using to see the avg DTU %:

SELECT end_time, 
  (SELECT Max(v)  
   FROM (VALUES (avg_cpu_percent), (avg_data_io_percent), (avg_log_write_percent)) AS  
   value(v)) AS [avg_DTU_percent] 
FROM sys.dm_db_resource_stats; 

Here's some example data:

end_time avg_DTU_percent
2015-09-11 05:03:05.740 98.58
2015-09-11 05:02:50.710 97.86
2015-09-11 05:02:35.680 97.92
2015-09-11 05:02:20.680 89.62
2015-09-11 05:02:05.633 0.21
2015-09-11 05:01:50.490 42.20
2015-09-11 05:01:35.337 37.96
2015-09-11 05:01:20.290 66.99
2015-09-11 05:01:05.273 64.36
2015-09-11 05:00:49.417 96.67
2015-09-11 05:00:34.387 97.72

I pulled a list of all current connections and did KILL for all active SPID. We have a couple of scheduled jobs that run periodically, so disabled those.

And still the DTU% keeps spiking.

This is completely out of character for our daily use but is clearly obvious to our users.

I'm now not sure if this is some problem in our application, or an issue on that particular Azure database.

Davideedle,

   Are you still facing this issue? You can query the DMV dm_exec_sessions (https://msdn.microsoft.com/en-us/library/ms176013.aspx) to see if there are any active sessions left in your database. The DMV dm_exec_query_stats will provide the recent query execution details to pin down the query causing the spikes. If you don't see any active sessions/queries from your side and the DTU is still spiking, please send your server, region and database name to me at srinia<.>Microsoft<.>com and we will take a look.

thanks

Srini Acharya

Hi Srini

Yes the issue persists today. I will email you now.

Thanks

David

FYI

Ive tried dm_exec_sessions but Im not sure Im seeing anything useful. I have no long running cursors. Yesterday I tried stopping everything that could be connecting to the database, and did KILL on on SPIDs. But once I reactivated everything the problem returned.

Its like my DTU allowance is suddenly back to 5 or 10 instead of 100.

FYI

I have bitten the bullet to try and get through the weekend and bumped our level from S3 to P2 to see if this helps in the short term.

One thing that strikes me is I cannot get the usage charts to load for this db either in the old or new portal. It does load for my other databases. Have no idea if there is a connection to my actual problem, but strikes me as odd.