Problem with Azure Storage Account (Used Capacity) Metrics

%3CLINGO-SUB%20id%3D%22lingo-sub-1147861%22%20slang%3D%22en-US%22%3EProblem%20with%20Azure%20Storage%20Account%20(Used%20Capacity)%20Metrics%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-1147861%22%20slang%3D%22en-US%22%3E%3CP%3EI%20have%20a%20bit%20of%20an%20odd%20issue%20in%20that%20when%20looking%20at%20the%20used%20Capacity%20metrics%20on%20my%20storage%20account%20it%20shows%20a%20wildly%20different%20amount%20from%20what%20is%20actually%20in%20use.%3CBR%20%2F%3EWe%20have%20had%20a%20storage%20account%20throttled%20as%20the%20capacity%20went%20over%20the%20maximum%20capacity%20of%20a%20storage%20account%20which%20we%20missed%20due%20to%20the%20Metric%20being%20incorrect.%3CBR%20%2F%3EThis%20is%20storage%20account%20hosts%20multiple%20unmanaged%20VHDs%20configured%20as%20Premium%20SSD%20disk.%3CBR%20%2F%3E%3CBR%20%2F%3EIs%20there%20a%20known%20issue%20or%20a%20suggested%20way%20to%20work%20around%20this%20%3F%3CBR%20%2F%3EWhere%20could%20I%20pull%20an%20accurate%20metric%20for%20the%20actual%20used%20amount%20of%20storage%20on%20the%20storage%20account%3F%3CBR%20%2F%3EAm%20I%20just%20missing%20something%20here%2C%20is%20there%20another%20Metric%20I%20should%20be%20looking%20at%20that%20might%20not%20be%20available%20via%20the%20GUI%20%3F%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-LABS%20id%3D%22lingo-labs-1147861%22%20slang%3D%22en-US%22%3E%3CLINGO-LABEL%3EAnalytics%3C%2FLINGO-LABEL%3E%3CLINGO-LABEL%3EAzure%3C%2FLINGO-LABEL%3E%3CLINGO-LABEL%3EAzure%20Essentials%3C%2FLINGO-LABEL%3E%3CLINGO-LABEL%3EAzure%20Resource%20Management%3C%2FLINGO-LABEL%3E%3C%2FLINGO-LABS%3E%3CLINGO-SUB%20id%3D%22lingo-sub-1149890%22%20slang%3D%22en-US%22%3ERe%3A%20Problem%20with%20Azure%20Storage%20Account%20(Used%20Capacity)%20Metrics%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-1149890%22%20slang%3D%22en-US%22%3E%3CP%3E%3CA%20href%3D%22https%3A%2F%2Ftechcommunity.microsoft.com%2Ft5%2Fuser%2Fviewprofilepage%2Fuser-id%2F453722%22%20target%3D%22_blank%22%3E%40hspinto%3C%2FA%3E%26nbsp%3BI%20have%20this%20problem%20with%20all%20the%20storage%20accounts%20on%20my%20tenant%20(~15).%20I%20normally%20split%20the%20load%20of%20the%20VMs%20are%20per%20best%20practice.%3CBR%20%2F%3EThrottling%20can%20be%20seen%20by%20looking%20under%20%22Resource%20Health%22%20for%20your%20specific%20storage%20accounts%20and%20reviewing%20the%20Health%20History's%20%22Health%20Events%22.%3CBR%20%2F%3E%3CBR%20%2F%3EIn%20terms%20of%20difference%20of%20%22%3CSPAN%3E%3CEM%3Ereal%20capacity%20usage%20value%20vs.%20the%20value%20observed%20in%20metrics%3C%2FEM%3E%22%20it%20looks%20something%20like%20this%20(Please%20view%20below%20table)%3C%2FSPAN%3E%3CBR%20%2F%3E%3C%2FP%3E%3CTABLE%20border%3D%221%22%3E%3CTBODY%3E%3CTR%3E%3CTD%3EStorage%20Account%20Name%3C%2FTD%3E%3CTD%3EMetric%20%22Used%20Capacity%22%3C%2FTD%3E%3CTD%3EReal%20Used%20Capacity%3C%2FTD%3E%3C%2FTR%3E%3CTR%3E%3CTD%3EStrgAcc1%3C%2FTD%3E%3CTD%3E41%20MiB%3C%2FTD%3E%3CTD%3E7000%20GiB%3C%2FTD%3E%3C%2FTR%3E%3CTR%3E%3CTD%3EStrgAcc2%3C%2FTD%3E%3CTD%3E150%20GiB%3C%2FTD%3E%3CTD%3E300%20GiB%3C%2FTD%3E%3C%2FTR%3E%3CTR%3E%3CTD%3EStrgAcc3%3C%2FTD%3E%3CTD%3E1%20GiB%3C%2FTD%3E%3CTD%3E4000%20GiB%3C%2FTD%3E%3C%2FTR%3E%3C%2FTBODY%3E%3C%2FTABLE%3E%3CP%3ESome%20things%20I've%20already%20considered%3C%2FP%3E%3CUL%3E%3CLI%3EMetrics%20%22Used%20Capacity%22%20does%20not%20match%20up%20to%3CUL%3E%3CLI%3ESize%20of%20non-VHDs%3C%2FLI%3E%3CLI%3ESize%20of%20any%20singular%20page%20blobs%3C%2FLI%3E%3CLI%3ESize%20of%20some%20odd%20combination%20of%20page%20blobs%3C%2FLI%3E%3C%2FUL%3E%3C%2FLI%3E%3C%2FUL%3E%3CP%3EThe%20%22Used%20Capacity%22%20does%20not%20seem%20to%20have%20any%20bearing%20on%20reality.%3CBR%20%2F%3EIn%20the%20case%20of%20say%20StrgAcc3%20there%20are%208%20page%20blobs%20and%20all%20are%20VHDs%20(Un-managed%20disks)%20and%20no%20one%20VHD%20matches%20the%20size%20of%20the%20%22Used%20Metric%22%3C%2FP%3E%3C%2FLINGO-BODY%3E%3CLINGO-SUB%20id%3D%22lingo-sub-1149408%22%20slang%3D%22en-US%22%3ERe%3A%20Problem%20with%20Azure%20Storage%20Account%20(Used%20Capacity)%20Metrics%3C%2FLINGO-SUB%3E%3CLINGO-BODY%20id%3D%22lingo-body-1149408%22%20slang%3D%22en-US%22%3E%3CP%3E%3CA%20href%3D%22https%3A%2F%2Ftechcommunity.microsoft.com%2Ft5%2Fuser%2Fviewprofilepage%2Fuser-id%2F492419%22%20target%3D%22_blank%22%3E%40Yuukan%3C%2FA%3E%2C%26nbsp%3Bcan%20you%20clarify%20what%20you%20mean%20by%20%22throttling%22%20and%20was%20the%20real%20capacity%20usage%20value%20vs.%20the%20value%20observed%20in%20metrics%3F%20400%20GiB%20of%20usage%20in%20a%20Storage%20Account%20is%20almost%20nothing%20compared%20to%20the%20Storage%20Account%20limits%20(in%20the%20order%20of%20TiB%20or%20even%20PiB).%20See%20%3CA%20href%3D%22https%3A%2F%2Fdocs.microsoft.com%2Fen-us%2Fazure%2Fstorage%2Fblobs%2Fscalability-targets%22%20target%3D%22_self%22%20rel%3D%22noopener%20noreferrer%20noopener%20noreferrer%22%3Escalability%20targets%20documentation%3C%2FA%3E.%3C%2FP%3E%0A%3CP%3E%26nbsp%3B%3C%2FP%3E%0A%3CP%3EBy%20the%20way%2C%20if%20your%20VMs%20are%20production%20ones%2C%20it%20is%20a%20bad%20idea%20to%20have%20multiple%20VMs%20sharing%20the%20same%20storage%20account%20-%20single%20point%20of%20failure.%26nbsp%3B%3C%2FP%3E%3C%2FLINGO-BODY%3E
Highlighted
Occasional Contributor

I have a bit of an odd issue in that when looking at the used Capacity metrics on my storage account it shows a wildly different amount from what is actually in use.
We have had a storage account throttled as the capacity went over the maximum capacity of a storage account which we missed due to the Metric being incorrect.
This is storage account hosts multiple unmanaged VHDs configured as Premium SSD disk.

Is there a known issue or a suggested way to work around this ?
Where could I pull an accurate metric for the actual used amount of storage on the storage account?
Am I just missing something here, is there another Metric I should be looking at that might not be available via the GUI ?

5 Replies
Highlighted

@Yuukan, can you clarify what you mean by "throttling" and was the real capacity usage value vs. the value observed in metrics? 400 GiB of usage in a Storage Account is almost nothing compared to the Storage Account limits (in the order of TiB or even PiB). See scalability targets documentation.

 

By the way, if your VMs are production ones, it is a bad idea to have multiple VMs sharing the same storage account - single point of failure. 

Highlighted

@hspinto I have this problem with all the storage accounts on my tenant (~15). I normally split the load of the VMs are per best practice.
Throttling can be seen by looking under "Resource Health" for your specific storage accounts and reviewing the Health History's "Health Events".

In terms of difference of "real capacity usage value vs. the value observed in metrics" it looks something like this (Please view below table)

Storage Account NameMetric "Used Capacity"Real Used Capacity
StrgAcc141 MiB7000 GiB
StrgAcc2150 GiB300 GiB
StrgAcc31 GiB4000 GiB

Some things I've already considered

  • Metrics "Used Capacity" does not match up to
    • Size of non-VHDs
    • Size of any singular page blobs
    • Size of some odd combination of page blobs

The "Used Capacity" does not seem to have any bearing on reality.
In the case of say StrgAcc3 there are 8 page blobs and all are VHDs (Un-managed disks) and no one VHD matches the size of the "Used Metric"

Highlighted

@Yuukan, bear in mind that despite having a 127 GB unmanaged disk attached to a VM, if that disk is consuming only 15 GB of space, Azure Storage metrics will only report 15 GB of used capacity for that disk. Can you check if that's the case?

 

You said that you have balanced your VMs across multiple storage accounts. That's fine, but if you move to Managed Disks, then you won't have to worry about disk placement anymore.  

Highlighted
I can confirm that the “Used Capacity” is not in line with the real capacity used within the VMs.
Some examples I’ve checked were for some DB and File servers where the Reported used capacity was only a few hundred (200-300) GiB but confirmed that actual data stored within multiple VHDs equates too 1 1/2 TiB
Highlighted

Dear @Yuukan, after what you described, I advise you to open a support ticket to get a better clarification. I don't see a good explanation for that metrics behaviour. I am sorry for not having been able to provide you with a solution.