Edit

Elastic SAN metrics

Azure provides metrics in the Azure portal that give you insight into your Elastic SAN resources. This article provides definitions of specific metrics you can select to monitor, their recommended use, how to view them at SAN level, and how to calculate IOPS and throughput.

View Elastic SAN metrics

  1. In the Azure portal, open your Elastic SAN resource.

  2. Under Monitoring, select Metrics.

  3. Select the Metric namespace and Metric.

  4. Select an aggregation.

Note

You can currently view Elastic SAN metrics only at the SAN resource level. Volume group and volume-level performance metrics aren't currently available for Elastic SAN. When you investigate a workload issue, use SAN-level metrics together with workload, VM, or operating system monitoring to isolate the source of the issue.

Available metrics

Elastic SAN metrics are grouped into two categories:

  • Capacity metrics describe capacity configured or allocated for the SAN.

  • Transaction metrics describe availability, requests, data transfer, and request latency.

Capacity metrics definitions

The following metrics are currently available for your Elastic SAN resource. You can configure and view them in the Azure portal:

Metric Definition Recommended use
Provisioned Base The baseline capacity you configure for the Elastic SAN when you create it. This capacity includes the initial reserved SAN capacity before you add any extra expansion capacity. Understand the base capacity that contributes to the SAN’s provisioned performance.
Provisioned Capacity The total currently provisioned SAN capacity. This capacity includes the original Provisioned Base plus any extra capacity expansions you add. Monitor the SAN’s configured capacity and capacity growth.
Provisioned Size The total allocated size of all volumes you provision within the SAN. Unlike Provisioned Capacity, which represents the total SAN capacity configured and available at the SAN level, Provisioned Size reflects how much of that SAN capacity you already assigned to volumes. Compare provisioned size with other SAN capacity metrics.
Snapshot Size The total snapshot size for the provisioned size under the SAN. Track snapshot capacity separately from volume capacity.
Total Provisioned Volume Capacity The aggregate logical footprint of provisioned volumes plus snapshots within the SAN. Understand how much capacity is provisioned to volumes.

Capacity metrics: Hierarchy and relationships

Before you monitor capacity metrics, it's important to understand how they relate to each other:

Provisioned Base
└─ Provisioned Capacity (base capacity + any purchased expansions)
 └─ Provisioned Size (sum of all provisioned volume sizes within the SAN)
  └─ Total Provisioned Volume Capacity (total allocated volume capacity)
  • You set Provisioned Base when you create the SAN. Expanding the SAN increases Provisioned Capacity, not the base.
  • Provisioned Size can never exceed Provisioned Capacity. If it does, volume creation fails.
  • Total Provisioned Volume Capacity is the total provisioned and allocated volume capacity, not the actual written usage. For example, provisioning a 10 GiB volume can result in 10 GiB of Total Provisioned Volume Capacity being reported, even if the volume contains minimal data.
  • Snapshot Size accumulates separately and isn't included in the Total Provisioned Volume Capacity. Monitor both together for total billed storage.

Transaction metrics definitions

Transaction metrics represent the total number of I/O requests that the Elastic SAN service processes. Transactions include successful requests, failed requests, and requests that return errors.

Metric Unit Definition Recommended use
Availability Percent The percentage of availability for the storage service or the specified API operation. Monitor service availability and configure availability alerts.
Transactions Count The number of requests made to a storage service or the specified API operation. This number includes successful and failed requests, as well as requests that produced errors. Understand I/O activity and calculate average IOPS over a selected interval.
E2E Latency Milliseconds The average end-to-end latency of successful requests made to the resource or the specified API operation. Understand latency across the complete request path.
Server Latency Milliseconds The average time used to process a successful request. This value doesn't include the network latency specified in E2E Latency. Compare with E2E Latency to determine whether latency is inside or outside service processing.
Ingress Bytes The amount of ingress data. This number includes ingress to the resource from external clients as well as ingress within Azure. Monitor write-oriented data transfer and calculate throughput.
Egress Bytes The amount of egress data. This number includes egress from the resource to external clients as well as egress within Azure. Monitor read-oriented data transfer and calculate throughput. This metric doesn't represent billable network egress.

Important

Elastic SAN currently surfaces foundational telemetry metrics such as Transactions, Ingress, and Egress instead of direct IOPS and throughput metrics. Depending on the selected aggregation and granularity, you can use the following calculations to derive IOPS and MB/s values.

IOPS calculation:

  • Average IOPS = Total transactions in time bucket / Seconds in time bucket

  • Example: A one-minute bucket contains 120,000 transactions:

    120,000 / 60 = 2,000 average IOPS

To calculate IOPS in the Azure portal:

  1. Select Transactions.

  2. Set Aggregation to Total.

  3. Select one-minute time granularity.

  4. Divide each displayed value by 60.

Important

This calculation produces the average rate for the selected time bucket. It doesn't show an instantaneous rate.

Throughput calculation:

  • Average total throughput (MB/s) = (Ingress bytes + Egress bytes) / Seconds in time bucket / 1,000,000

For separate read and write views:

  • Average write throughput (MB/s) = Ingress bytes / Seconds in time bucket / 1,000,000
  • Average read throughput (MB/s) = Egress bytes / Seconds in time bucket / 1,000,000

Important

In the context of Elastic SAN, Ingress represents data written to the SAN (coming from your VM/client), and Egress represents data read from the SAN. These definitions aren't the same as conventional Azure network ingress/egress terminology.

Resource logging

You can configure the diagnostic settings of your elastic SAN to send Azure platform logs and metrics to different destinations. Currently, there's one log configuration:

  • Transactions - Every transaction log that the resource offers.

Audit logs are each resource provider's attempt to provide the most relevant audit data. However, these logs might not be considered sufficient from an auditing standards perspective.

Available log categories:

  • Write Success Requests
  • Write Failed Requests
  • Read Success Requests
  • Read Failed Requests
  • Persistent Reservation Requests
  • SendTargets Requests

Monitor workload performance

Monitor Availability

The Availability metric can be useful for viewing any visible problems from either an application or user perspective.

When you use this metric with Azure Elastic SAN, you should use the Average aggregation. By using Average, you see what percentage of your requests are experiencing errors and if they're within Elastic SAN's SLA. For worst and best case availability scenarios, you should use the Min and Max aggregation. For example, you can use Min availability for incident alerting.

How to create an alert for Availability < 99.9%

  1. Open the Create an alert rule dialog box. For more information, see Create or edit an alert rule.

  2. In the Scope tab, select your Elastic SAN resource.

  3. In the Condition tab, select the Availability metric.

  4. In the Alert logic tab, select the following attribute variable values from the drop-down menu:

    Field Description
    Threshold Static
    Aggregation type Average
    Operator Less than
    Threshold value 99.9

  5. In the When to evaluate tab, select the following variable values from the drop-down menu:

    Field Description
    Check every 5 minutes
    Lookback period 1 hour

  6. Select Next to go to the Actions tab and add an action group (email, SMS, and so on) to the alert. You can select an existing action group or create a new action group.

  7. Select Next to go to the Details tab and fill in the details of the alert such as the alert name, description, and severity.

  8. Select Review + create to create the alert.

Next steps

  • Last updated on