Working in Sizer with Collector Data

Once you have gathered the workload requirements using Collector, you can generate the Nutanix solutions that are best suited for your workload needs by either

  1. Export the Collector data to Sizer via the “Export to Sizer” option within the Collector Portal or
  2. Import the Collector data in Sizer via the “Import Workloads” option within the Sizer Scenario

Note: Apart from Nutanix Collector, Sizer supports importing output from RVTools and Nutanix Insights. The concepts and parameters remain the same irrespective of the import source. 

There are various options while feeding the Collector data to Sizer. This help page will describe how and when to use the different available options.

In the following section, we will be exporting the data from the Collector Portal to Sizer. To export the data to Sizer, navigate to the desired Project in the Collector Portal, click Export, and select the Export to Sizer option as shown in Figure 1 below.

Figure 1: Exporting Collector Project to Sizer

On clicking “Export to Sizer” you will be shown a pop-up dialog asking for Sizer scenario details as seen in Figure 2 below.

Figure 2: Specifying Scenario Details

In the first step, you can specify the Scenario Name and Account name. Optionally, you can specify the Opportunity details are well. The default name of the scenario is the same as the Collector Project name.

Note: Account and Opportunity detail fields are limited to Nutanix Employees and not visible to partners. 

In the second step, the user can specify all the “Sizing Preferences” as seen in Figure 3 below

Figure 3: Specifying Sizing Preferences

At a very high level, the sizing preferences are based on the number of VMs and their requirements – the compute, memory & storage.

Now, let’s start with the “VMs” option – Powered ON only vs Both powered ON and powered OFF. Nutanix Collector gathers configuration and performance data of all the VMs managed by the IP address specified while establishing the connection. If there are VMs that you don’t wish to size, you can optionally turn them off using the “VM List” tab before exporting to Sizer.  In most scenarios, we recommend the users to choose the “Powered ON only” option but you do have an option to select “Both powered ON and powered OFF” VMs in case you plan to run all the discovered VMs on the Nutanix solution.

Figure 4: Export VMs – Powered ON only vs Both powered ON and powered OFF

Warning: When the data gathered is that of a VDI workload, we highly recommend our users carefully look at the CPU utilization values of the powered-off VMs under the “VM List” tab. The VMs may be powered off when Collector gathered the data but these VMs may have been running intermittently and might be consuming resources. If you miss out on exporting these VMs, you might end up under-sizing the solution. 

Next, let us look at the “Capacity” option – Consumed vs Provisioned. This option lets the users decide if they would like to size for storage based on the actual consumed storage or the storage that is provisioned for each VM. The selection purely depends on the actual storage utilization value and the future capacity needs.

Figure 5: Export Storage - Consumed vs Provisioned

Figure 5: Export Storage – Consumed vs Provisioned

If the current environment is over-provisioned on storage, we recommend the users go with the “Consumed” option.  Don’t worry about running out of storage if you go by the “Consumed” option, Sizer does take care of it and also ensures the recommended solution is capable of handling node failure.  Alternatively, you can always go with the “Provisioned” storage values if you are sure that the allocated storage is required or you want to take a more conservative approach.

Note: When sizing only powered-ON VMs, do ensure to size the storage of powered-OFF VMs. The VMs are powered OFF but the data of powered OFF VMs persists and needs to be accounted for during sizing. 

Now, the most critical part, sizing based on Configuration vs Performance. This option will decide the compute or cores within the solution. Again the selection of this option depends on the CPU utilization values. When we look at the data collected, we see a lot of existing solutions are oversized on the compute or cores. This is usually due to the bad job done during requirement gathering or a very conservative approach taken to avoid CSAT issues.

Figure 6: CPU Utilization under Cluster Summary

The cluster summary gives an overall CPU utilization of the cluster and the “Provisioning Status” donut under VMs Summary helps to identify the VMs that are over-provisioned or under-provisioned.

Figure 7: Provisioning Status under VMs Summary

The above two charts should give the users a fair idea and help users to decide on which option to go with when it comes to sizing CPU cores or compute. If the CPU utilization value is very low, we recommend going with “Performance” based sizing. When using the performance-based option, we highly recommend using 95th percentile CPU utilization values to ensure the workload demands are met all the time.

Figure 8: Export Workload CPU options – Configuration vs Performance

Warning: Selecting performance based on Median or Average may result in under-sizing the solution.

When using 95th percentile-based CPU utilization values, Sizer does factor in an additional 20% of compute requirements to ensure the workload spikes are met. The end Sizer solution has enough cores within the solution to ensure you are not short on CPU cores and the solution can handle node failures as per desired resiliency levels (that can be tweaked in Sizer).

The next preference is the “Create Workloads by” option – you can either choose Profile or VM-based workloads. Ever since the launch of “Export to Sizer” we supported “Profile” based workloads. This approach would categorize each collector discovered VM under predefined VM buckets in Sizer – XS, S, M, L, XL.

Figure 9: Export Workload – Profile-based vs VMs based

Beginning in 2021, we have started supporting “per VM” based workloads, this results in a much more accurately defined workload feed to Sizer.  We highly recommend the “per VM” based option to our users. This option currently has a limitation and supports only up to 500 VMs. If you are trying to export more than 500 VMs you will have to go with “Profile” based workloads.

Retaining VM to Cluster Mapping is another handy preference – the option is unchecked by default. By default, exporting to Sizer consolidates all the VMs in one cluster (Cluster-1), this is good if the customer wants to consolidate the VMs within one cluster. But if the users want to recreate the same cluster and VM mappings in the Nutanix solution, you can always use “Retain VM to Cluster Mapping” to ensure AS-IS VM to cluster mapping is retained in the Nutanix Sizer solution.

The final option is the  Flash percentage – the Flash percentage option allows the user to control the Flash storage requirements.

Figure 10: Specifying workload Flash Requirement

By default, the storage requirements are split into 90:10, 90% HDD requirements, and 10% flash requirements. Depending on the workload nature, the users can tweak this percentage as desired. If the users prefer an all-flash solution, the slider value can be set to 100%.

Now, we are all set to hit the “Done” button to export to Sizer. On successfully exporting to Sizer, you can see the newly created Scenario URL in the banner as shown in Figure 10 below.

Figure 11: Banner showing Sizer Scenario URL

Successful export will also generate a CollectorToolsSummaryReport.xlsx with all the details of the VM data.

Hopefully, this help page provides you with more clarity on the options to choose when exporting Collector data to Sizer. By the way, you will also see the same options when you are trying to “Import” workloads in Sizer either via Collector or RVTools data.

If you have any queries feel free to reach out to Collector Support via collector@nutanix.com or alternatively you reach the Collector team via the Nutanix Community Page – Sizer Configuration Estimator

For more information please refer to Nutanix Collector User Guide, Security Guide, Release Notes & Nutanix Collector Portal User Guide

For any other queries related to Nutanix Collector, please refer to Collector FAQs

Collector 3.3

While the previous release of the product added support for configuration collection of multi-node Hyper-V clusters, Collector 3.3 has two major updates:

1) Option to collect performance metrics of Hyper-V clusters

2) Support for single-node Hyper-V cluster

This means with Collector 3.3 you can now Size your Hyper-V clusters either based on configuration or performance data.

The release also has fixes for the Hyper-V connection issues observed on systems running in languages other than English. The new release also provides an option to connect to the Hyper-V cluster remotely via HTTPS port 5986 or HTTP port 5985.

Collecting configuration & performance metrics of your Hyper-V cluster

Hyper-V connection modes

There are two modes via which you can collect data of your Hyper-V clusters:

1) Hyper-V (Local) – Recommended

2) Hyper-V (Remote)

It is highly recommended to use the Hyper-V (Local) mode to connect to your Hyper-V clusters. This would avoid any connection issues due to Firewall settings or proxy servers.

If your Hyper-V hosts are running Windows Core Operating system you can connect to the cluster using Hyper-V (Remote) mode.

Prerequisites for connecting to Hyper-V cluster

Multi-node Hyper-V Cluster

1) The credentials used should be of the Domain Admins group user

2) Cluster hosts should be part of a Microsoft Failover Cluster Manager

3) PowerShell 5.1 or later

Single-node Hyper-V Cluster

1) The credentials should be of a Local Admin user

2) PowerShell 5.1 or later

Collecting performance metrics of a Hyper-V cluster

In the case of the Hyper-V cluster, the performance data is not collected unless it is explicitly initiated by the user. The performance data collection is live data collection as opposed to ESXi or AHV hypervisor environments.

You can collect performance data ranging from anywhere from 1 day to 7 days. The frequency of data collection is automatically decided based on the duration of the performance collection.

The below table provides details of the frequency and number of data points that are collected for the supported durations:

Duration of Performance Collection Frequency of Data Collection Data points over the duration
1 day 5 minutes 288
3 days 10 minutes 432
5 days 20 minutes 360
7 days 30 minutes 336

Once the data points are collected, the Cluster Summary, Host Summary & VM Summary charts will automatically start populating the performance charts & data. Below is a sample screenshot of the VM Summary screen:

Demo of Collector 3.3

Connecting to Hyper-V cluster

CVM (Cores, Memory, HDD, SSD)

CVM Cores & Memory Overheads

The CVM (Controller VM) CPU core and memory requirements in Nutanix environments are determined by a couple of key parameters:

Total Capacity

  • The overall storage capacity of the node influences the number of CPU cores and the amount of memory allocated to the CVM.
  • Higher capacity nodes require more resources to manage storage operations efficiently.

Recovery Point Objective (RPO)

  • The RPO, which defines the acceptable amount of data loss in case of a failure, impacts the CVM resource allocation.
  • Lower RPOs require more CPU cores and memory to ensure data replication is quick and accurate.

For the actual resource requirements refer to HCI Node Capacity & CVM Resource Requirements.

CVM Storage Overheads

Key Parameters for CVM Storage Overhead:

  • Nutanix Boot + Home: 280 GB across 2 drives, always on Tier-1
  • OpLog: 600 GB, Tier-1
  • Hades Reservation: 100 GB per disk, irrespective of disk type
  • Cassandra: 3% of all drives, are always on Tier-1 storage
  • Curator: 2% of all drives, are always on Tier-2 storage
  • Ext4 Formatting: Varies by storage type (HDD, SSD, NVMe)

CVM Storage Overhead Summary:

Nutanix Boot + Home:

  • Overhead: 280 GB across 2 drives
  • Tier: Always on Tier-1 Storage

OpLog:

  • Overhead: 600 GB
  • Tier: Always on Tier-1 Storage

    Hades Reservation:

  • Overhead: 100 GB per disk
  • Tier: Irrespective of disk type

    Cassandra:

  • Overhead: 3% of all drives
  • Tier: Always on Tier-1 Storage

Curator:

  • Overhead: 2% of all drives
  • Tier: Always on Tier-2 Storage

    Ext4 Formatting:

  • HDD: 1.78% of drive size
  • SSD & NVMe: 2.6% of drive size

The overheads can be seen in Sizer by hovering over the info icon as seen below:

Continue reading “CVM (Cores, Memory, HDD, SSD)”