Azure VM Disk IOPs report with Disk details

Hi Logan
I just wanted to check if the above provided information worked for you or if you need any further assistance?

Please feel free to let us know, as we are always here to help whenever you need us.

Additionally, if you have a moment, please do upvote it if the information provided has been helpful. This can be beneficial to other community members and would be greatly appreciated.

Thank you.

Hi Logan,

I would like to check if the issue has been resolved or if you need any assistance. Please feel free to let us know, as we are always here to help.

Please do not forget to “upvote it” wherever the information provided helps you, this can be beneficial to other community members.it would be greatly appreciated and helpful to others.

Thanks

Hi Pranay Reddy Madireddy,

The premium disk filter was already applied in the arg query itself, so I just added only the summarize function as per your recommendation, but it didn't work and got below error. Were you able to execute a query with your suggestion?

Hi Logan,

I reviewed your error, and it looks like the issue is with how arg_max() it is being used. In Kusto, the first argument to arg_max() must be a scalar value like TimeGenerated, Val, or any sortable primitive — not a structured or nested type.

If you're trying to get the latest entry per disk, you can modify the line like this:

    summarize arg_max(TimeGenerated, *) by _ResourceId

This will return the most recent record for each _ResourceId, assuming TimeGenerated is present in your dataset (which it should be for metric data). If TimeGenerated isn't available in that scope, you could try using another numeric field like Val if appropriate.

Please let us know if it was helpful, and feel free to reach out if you have any further queries.

Hi Vinay B,

Appreciate your response here. I'm still getting Disk letters are duplicated multiple times and attached is the code and the results for a single VM.

// Summarize the disk data (only one row per disk/resource)
let DiskData = 
    arg("").Resources
    | where type == "microsoft.compute/disks"
    | project
        diskName = name,
        diskSizeGB = properties.diskSizeGB,
        resourceGroup,
        location,
        OSType = properties.osType,
        DiskType = sku.name,
        _ResourceId = tolower(managedBy),
        Throughput = properties.diskMBpsReadWrite,
        IOPSReadWrite = properties.diskIOPSReadWrite,
        AppName = tags.applicationName
    | where DiskType contains "Premium";
// Summarize InsightsMetrics data (only one row per Computer/Disk combination)
let ReadIops = 
    InsightsMetrics
    | where Origin == "vm.azm.ms"
        and Namespace == "LogicalDisk"
    | where Name == "ReadsPerSecond"
    | extend Disk=tostring(todynamic(Tags)["vm.azm.ms/mountId"])
    | summarize arg_max(TimeGenerated, *) by Computer, Disk, Val, _ResourceId
    | project
        TimeGenerated,
        Computer,
        Disk,
        ReadIOPSMax = Val,
        ReadIOPSResID = _ResourceId;
let ARGINS = DiskData
    | join kind=inner (ReadIops) on $left._ResourceId == $right.ReadIOPSResID;
    ARGINS
    | project Computer, diskName, diskSizeGB, Disk, ReadIOPSMax, Throughput, IOPSReadWrite

Hi Logan,

Thanks for the update and for sharing the results. The reason you're still seeing duplicate drive letters (Disk) is because arg_max() is currently grouping by Val, which varies across time and effectively prevents deduplication.

To ensure only one row per Computer + Disk, we should remove Val from the summarize grouping keys. Here's the corrected version.

let ReadIops = 
    InsightsMetrics
    | where Origin == "vm.azm.ms"
        and Namespace == "LogicalDisk"
    | where Name == "ReadsPerSecond"
    | extend Disk = tostring(todynamic(Tags)["vm.azm.ms/mountId"])
    | summarize arg_max(TimeGenerated, *) by Computer, Disk, _ResourceId
    | project
        TimeGenerated,
        Computer,
        Disk,
        ReadIOPSMax = Val,
        ReadIOPSResID = _ResourceId;

This change ensures arg_max() picks only the most recent entry per disk, per VM, avoiding duplication caused by slight variations in Val. You can apply the same pattern to Write IOPS if needed.

Please let us know if it was helpful, and feel free to reach out if you have any further queries.

Hi @Vinay B,

I'm not sure it works because i keep getting the same duplicated results and incorrect value in IOPS.

    InsightsMetrics
    | where Origin == "vm.azm.ms"
        and Namespace == "LogicalDisk"
    | where Name == "ReadsPerSecond"
    | extend Disk = tostring(todynamic(Tags)["vm.azm.ms/mountId"])
    | summarize arg_max(Val, *) by Computer, Disk, _ResourceId

@Clive Watson,

It's still the same. I'm even thinking whether this is achievable or not. Tried lot of possibilities and getting same duplicated results in different ways not the actual results.

This is combined query result:

It's from InsightMetrics without ARG query join:

@Logan I tried to reshape your query a bit. The one I tried (see below) is not showing duplicates. I suspect duplicate disk entries were caused by replicas not excluded from the ARG query. Have a look at my version and see if that solves your problem giving you the desired results:

// Summarize the disk data (only one row per disk/resource)
let DiskData = arg("").Resources
    | where type == "microsoft.compute/disks"
    | where sku.name contains "Premium"
    | extend
        diskName = name,
        diskSizeGB = properties.diskSizeGB,
        resourceGroup,
        location,
        OSType = properties.osType,
        DiskType = sku.name,
        _ResourceId = tolower(managedBy),
        Throughput = properties.diskMBpsReadWrite,
        IOPSReadWrite = properties.diskIOPSReadWrite,
        AppName = tags.applicationName
    | where isnotempty(_ResourceId)
    | where isnotempty(OSType)
    | project diskName, diskSizeGB, resourceGroup, location, OSType, DiskType, _ResourceId, Throughput, IOPSReadWrite, AppName;
// Summarize InsightsMetrics data (only one row per Computer/Disk combination)
InsightsMetrics
| where Name in ('FreeSpaceMB', 'ReadsPerSecond', 'WritesPerSecond')
| extend Disk = tostring(parse_json(Tags).['vm.azm.ms/mountId'])
| summarize arg_max(TimeGenerated, *) by _ResourceId, Disk, Name
| extend _ResourceId = tolower(_ResourceId)
| extend DisksizeinGB = iif(Name=='FreeSpaceMB', round(toreal(parse_json(Tags).['vm.azm.ms/diskSizeMB'])/1024.0,2), 0.0)
| extend AvailablespaceinGB = iif(Name=='FreeSpaceMB', round(Val/1024.0,2), 0.0)
| extend UsedSpaceinGB = iif(Name=='FreeSpaceMB', round(DisksizeinGB-AvailablespaceinGB, 2), 0.0)
| extend DiskUtilizationPercentage = iif(Name=='FreeSpaceMB', round((UsedSpaceinGB / DisksizeinGB) * 100, 2), 0.0)
| join hint.remote=left kind=innerunique DiskData on _ResourceId
| project TimeGenerated, Computer, _ResourceId, diskName, Disk, Name, Val, DisksizeinGB, AvailablespaceinGB, UsedSpaceinGB, DiskUtilizationPercentage, diskSizeGB, resourceGroup, OSType, DiskType, Throughput, IOPSReadWrite, AppName