Requesting clarification on Azure Advisor VM Rightsizing recommendations

Anil Kumar 525 Reputation points
2023-09-23T07:08:04.2366667+00:00

Hello,

We are analyzing VM Rightsizing recommendations generated by Azure Advisor with 90 days look back period and 100% CPU utilization in its configurations rules.

Could you please clarify following questions on ?

  1. Does Azure Advisor take into consideration the count of disk attached to a VM while generating Rightsizing recommendation? If 17 disk are attached at present, will Azure recommend only VM sizes who can support at least 17 disks.
  2. Does Azure Advisor takes into consideration the storage throughput: IOPS/MBps and Max uncached disk throughput: IOPS/MBps of VM target to ensure if equals to or greater than the max IOPS/MBps of VM current size in Rightsizing recommendation? If VM current size can support 25600 IOPS but max IOPS on that VM is only 6000 - Would Azure Advisor recommend VM sies which can support at least 6000 IOPS?
  3. Does Azure Advisor takes into consideration the max or P95 max CPU usage of VMs while generating Rightsizing recommendation? Below article mentions:
    1. Recommendation criteria include CPU, Memory and Outbound Network utilization.
      1. Metrics are sampled every 30 seconds, aggregated to 1 minute and then further aggregated to 30 minutes (taking the max of average values while aggregating to 30 minutes).

Not understanding how Azure Advisor can make right recommendations considering only AVG metrics utilizations.

I feel like executing Azure Advisor VM Rightsizing recommendations is very risky, are there any pointers that should we take into consideration to ensure target VM size is rightly sized.

https://learn.microsoft.com/en-us/azure/advisor/advisor-cost-recommendations

Applicate your Response. Thank you!!

Azure Virtual Machines
Azure Virtual Machines
An Azure service that is used to provision Windows and Linux virtual machines.
9,058 questions
{count} votes

Accepted answer
  1. vipullag-MSFT 26,492 Reputation points Moderator
    2023-09-28T07:10:10.26+00:00

    Hello Anil Kumar

    Sharing responses for your queries:

    -Does Azure Advisor takes into consideration the max or P95 max CPU usage of VMs while generating Rightsizing recommendation? Below article mentions:

    Recommendation criteria include CPU, Memory and Outbound Network utilization.

    Metrics are sampled every 30 seconds, aggregated to 1 minute and then further aggregated to 30 minutes (taking the max of average values while aggregating to 30 minutes).

    Yes, Azure Advisor does take into consideration the maximum CPU utilization of the VMs while generating Rightsizing recommendations. The article you mentioned is correct that the metrics are sampled every 30 seconds, aggregated to 1 minute, and then further aggregated to 30 minutes (taking the max of average values while aggregating to 30 minutes). This means that Azure Advisor considers both the average and maximum CPU utilization of the VM over a 30-minute period to determine the VM usage.

    The article you mentioned is also correct that the recommendation criteria include CPU, Memory, and Outbound Network utilization.

    -Also how VM usage can be determined based on AVG metrics data - on one VM, I see CPU reaching 80% or more couple of times in a day but its AVG CPU utilization is very less?

    Azure Advisor uses a combination of average and maximum values to determine VM usage. The metrics are sampled every 30 seconds, aggregated to 1 minute, and then further aggregated to 30 minutes (taking the max of average values while aggregating to 30 minutes). This means that Azure Advisor considers both the average and maximum CPU utilization of the VM over a 30-minute period to determine the VM usage.

    In your example, if the CPU reaches 80% or more a couple of times in a day but the average CPU utilization is very low, it is possible that the high CPU utilization is occurring during peak usage periods. Azure Advisor takes this into account by considering the maximum CPU utilization over a 30-minute period, which would capture the peak usage periods.

    I understand your concern about executing Azure Advisor VM Rightsizing recommendations based on average metrics utilization. While it is true that Azure Advisor uses average metrics utilization to generate recommendations, it also takes into account the maximum utilization over a 30-minute period. This means that Azure Advisor considers both the average and maximum CPU utilization of the VM over a 30-minute period to determine the VM usage.

    Hope this clarifies.

    1 person found this answer helpful.

1 additional answer

Sort by: Most helpful
  1. AirGordon 7,150 Reputation points
    2023-09-23T08:17:20.3933333+00:00

    My experience is that the Advisor recommendations are a good starting place, but do not take into account many of the variables.

    The docs do provide a comprehensive list of factors that are used in the recommendation: https://learn.microsoft.com/en-us/azure/advisor/advisor-cost-recommendations#resize-sku-recommendations

    I'd suggest that you leverage standard practices like making changes and testing in representative non-prod environments before making changes to production environments to then build confidence that the recommendation is correct.


Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.