Skip to main content
Pure Technical Services

Pure1 Manage - VM Analytics

Currently viewing public documentation. Please login to access the full scope of documentation.

Pure1 VM Analytics Requirements

  • OVA Collector
    • No Purity requirements to run OVA Collector
    • 4 Cores and 8GB Memory
    • Username/password - Unique read-only user specifically for collector
    • IP address of vCenter
    • Supported to be deployed on vCenter 6.5 U3+ (6.5.0.3000+)
    • Supported to collect from vCenter 5.5+
  • Must have phonehome enabled
  • VM Analytics supports VMFS, vVols, vSAN, and NFS datastores
  • VMware Tools 10.1+ must be installed to see VM Capacity metrics

Pure1 VM Analytics Highlights

Pure1 is here to help narrow down the troubleshooting steps in your virtualized environment. VM Analytics provides you with a visual representation of the I/O path from the VM all the way through to the FlashArray. Other tools and features guide you through identifying where an issue might be occurring in order to help eliminate potential candidates for a problem.

VM Analytics doesn’t only help when there’s a problem. The visualization allows you to identify which volumes and arrays particular applications are running on. This brings the whole environment into a more manageable domain. 

New:

Instructions

VM Analytics Collector

The Collector is a pre-packaged OVA appliance that sits at the heart of VM analytics. You can connect multiple vCenters to a single collector, so you only need to deploy one in your environment. However, you can deploy multiple collectors when necessary, for example, if all vCenters are not on the same network.

Collection times may be impacted by the number of VMs that are being collected. Typically, a collector can support 8,000 VMs within the default collection period, but this scale may be impacted by issues such as latency between the collector and vCenter.

The VM Analytics Collector has negligible impact on vCenter performance, so you can deploy the collector safely without impacting current workloads. We expect less than a 5% impact on CPU utilization while we are collecting metrics (by default every 10 minutes).

A correctly configured collector is required for the Analytics > VM Topology page to reflect the topology and metrics from your vSphere environment.

The On-Array Collector

The on-array collector is deprecated and is no longer supported for new installs. Please instead use the OVA-based collector.

The OVA Collector

New Features with Collector Version 3.0
New capacity metrics

Capacity data for VMs and datastores are collected and reported on the Pure1 Manage > Analytics >VM Topology page.
See Capacity Metrics.

New Controller Commands 
  • OVA deployment cleanup
    The puresetup vm-analytics-collector command completes the OVA installation. Used only if the OVA installation did not complete during the Deploy OVF Template... step in vSphere.
  • Collector registration
    Register the collector with Pure1 using the purevmanalytics register command and the authorization key from the Pure1 Manage the VM Topology > Collector Configuration page.
    purevmanalytics register <authorization_key>
  • Support bundle command
    Prepare logs to send to Pure Technical Services or to view log contents:
    puresupport bundle
  • New upgrade procedure (to upgrade to future collector releases after v3.0.0)
    Run the puresw list command to check for available OVA and collector updates.
    Run the puresw upgrade start vm-analytics-collector command to perform updates.
New Steps for v3.0.0 OVA collector creation, deployment, and configuration
  1. In Pure1 Manage > Analytics > VM Topology, create a new collector, copy its authorization key, and download the OVA collector or copy its link.
    See Create and Configure the Collector in Pure1 Manage.
  2. In vSphere, install the OVA collector using Deploy OVF Template... and power on the OVA VM.
    See Deploy the OVA Collector in vSphere.
  3. Log into the OVA VM.
    Change the pureuser password when prompted.
    If required, run puresetup vm-analytics-collector to complete the OVA deployment.
    Run the purevmanalytics register xxxx-xxxxx-xxxxx-xxxxx-xxxxxx-xxxx command.
    Run the purevmanalytics config import command to import the configuration of the previous (currently running) collector.
    See Configure the OVA VM.

 

OVA Collector Creation, Deployment, and Configuration
Create and Configure the Collector in Pure1 Manage

You download the OVA from pure1.purestorage.com.
To begin, on the Pure1 Manage > Analytics > VM Topology page, click the settings icon.

In the Collector Configuration page, you create a collector, name it, copy its authorization key, and either download the OVA collector or copy its link for use in vSphere:

  1. Under OVA Collectors, click Create Collector.

    clipboard_e6c14cb21c5ebf4a123e2f7a700d30cfd.png
  2. A name field open next to the Create Collector button. Give the collector a name (this can be any name -- it is not checked against the eventual collector VM name, though it should match for ease of management).
    After entering the name, again click Create Collector.

        clipboard_e9452b6257f47ae78e169c0e7c9098fad.png
  3. Your new collector appears at the top of the OVA Collectors table. Click the copy icon on its row to copy the collector's authorization key, which is required for OVA deployment on vSphere.
    coll-new-entry.png
  4. The download link for the OVA is at the bottom of the Collector Configuration page.


    Either download the OVA Collector to your local machine or copy this URL by right clicking > Copy Link Address, then enter the copied address in the vSphere Deploy OVF Template wizard.
    The download link is not unique and can be reused for multiple OVA installs.
Deploy the OVA Collector in vSphere

In vSphere, install the OVA Collector with the Deploy OVF Template wizard.

  1. Navigate to the host that you’d like to install this on, right-click on the host, and select Deploy OVF Template…

         
  2. Once you’re there, you can type in the OVA's URL (copied during Collector Configuration) or upload a local file to vCenter.
    • URL: Enter https://static.pure1.purestorage.com/vm-analytics-collector/purestorage-vma-collector_latest-signed.ova.
    • Local File: Select Local File, browse to the OVA file (that you downloaded previously), and follow the instructions on the screen.


       
  3. Next you’ll select a name and location for the VM and click Next

         

  4. Then select the compute resources for the VM and click Next

        
  5. Verify the information is correct and click Next

        
     
  6. Scroll down to read the entire license agreement. Accept the license agreement by clicking the I agree to all licenses checkbox.

        
  7. Here, you’ll probably want to change the Virtual Disk Format to Thin Provisioned. Then select which datastore you’d like and click Next

        
  8. Select the Destination Network.

        
        
  9. Finally, you have the customize template page. Here, set up your network settings. You’ll have the option to choose DHCP, or you can set up a static IP. Click Next, then click Finish and the OVA VM will deploy.

        
  10. The deployed OVA collector appears as a VM.
        deploy-ova_9d_power.png
  11. Power on the OVA VM.
Configure the Collector

Perform these steps after the OVA has been installed in vCenter.

You are required to change the password for the pureuser account when you first log in. Be sure to note your new password. If pureuser cannot log in, you will have to redeploy the OVA to gain access.

  1. Open an SSH connection to the collector using the OVA VM's DNS name or IP address displayed in vCenter.
  2. In the collector shell:
    1. On first login, you are prompted to change the pureuser password.
    2. Run the puresetup vm-analytics-collector command only if the OVA installation did not complete during the Deploy OVF Template... step in vSphere.

      If the deployment is incomplete, you see this message when logging into the OVA VM:

      There was an error while performing application setup:
        Error: Error installing application.
      Use `puresetup` to try again.

      if you see this message, run the puresetup vm-analytics-collector command.
      Example command and output:
      $ puresetup vm-analytics-collector
      Checking for available versions...

      Ready to install  3.0.0...
      Do you want to continue? [Y/n]: y

      Installing  3.0.0...
      Installation complete!
      The following new commands have been added:
        purevmanalyticscommon.py
        purevmanalytics

      Network settings may prevent puresetup from completing successfully. If puresetup fails repeatedly, contact Pure Technical Services.

      Again, only run puresetup if you are prompted to do so.
      The puresetup command is not available when the collector deployment is complete.

    3. Register the collector.
      Registration is a one-time step required only after the first install of a v3.0.0 collector.
      Registration requires the authorization key assigned the collector in the Pure1 Manage Collector Configuration page (see Create and configure the collector in Pure1 Manage).

      $ purevmanalytics register <authorization_key>
      Registration completed successfully.
    4. If you are upgrading from an off-array collector or from a v2.x collector, import the configuration of the previous collector.
      $ purevmanalytics config import --host <old_collector_host_name> --user <username>

      Notes:
      • old_collector_host_name is the IP address of the collector you are importing from.
      • username is the username on the collector you are importing from.
      • You are prompted for that user's password.
      • The import includes connections to the vSphere instances used by the v2.x collector.
      • Import is designed to bring the configuration of a v2.x collector to a v3.0.0 (or higher) collector.
        Import is not supported between v.3.0.0 (or higher) collectors.
    5. If you are not upgrading from a v2.x collector, connect the collector to the relevant vSphere instances.
      $ purevmanalytics connect --hostname <VCENTER> --username <USERNAME@domain.name>

      Notes:
      • VCENTER is the DNS name or IP address of the vSphere instance the collector will pull metrics from.
      • USERNAME@domain.nameis the login for the vSphere instance the collector will pull metrics from.
      • You are prompted for that user's password.
      • Run purevmanalytics connectonce for each vSphere instance the collector will pull metrics from.

After some delay, your metrics will be populated in the Pure1 Manage > Analytics > VM Topology page. The delay could be a couple of hours, depending on the size of the vCenter. The initial phonehome includes the entire topology, so it may take slightly longer. Subsequent phonehomes only track changes in order to reduce the size of the logs and transfer time.

Phonehome Settings

This information is provided to aid in configuring firewalls and proxies.

Port

Transport Protocol

Application Protocol

Direction

Additional Information

443

TCP

HTTPS

Outbound

For IPv4, use the following IP block and hostnames:

  • IP block 52.40.255.224/27 [1]

  • Hostnames:

    • *.cloud-support.purestorage.com

For IPv6, we currently do not support static IPs. Use the following hostnames:

  • *.cloud-support.purestorage.com

Please add all of the above to your whitelist.

[1] This is an IP CIDR block.  This refers to 32 IPs (52.40.255.224 - 52.40.255.255).  Most firewalls will accept the whole block, so it can be entered as one-line in most cases.

123 UDP UDP Outbound The NTP service uses this port to synchronize the OVA appliance time to the NTP servers: 208.81.1.244, 167.114.96.24, 209.115.181.102.

For collector upgrades, you will need to specifically whitelist addresses from:

nslookup deb.cloud-support.purestorage.com 

Information Collected

The collector gathers information every 600 seconds. The information collected includes the following.

Type of Information Objects
Topology VMs, hosts, virtual disks, and datastores
I/O metrics VMs, hosts, virtual disks, and datastores
CPU and memory metrics VMs and hosts
Capacity metrics VMs and datastores

 

Man Page - purevmanalytics 

purevmanalytics -h
usage: purevmanalytics [-h] [-v]
           {register,start,stop,restart,config,list,connect,disconnect,enable,disable,version,phonehome,stackdump,test}
           ...
 
Pure Storage VM Analytics Collector Configuration v3.0.0-20200925-3e0a406
 
positional arguments:
  {register,start,stop,restart,config,list,connect,disconnect,enable,disable,version,phonehome,stackdump,test}
    register            register authorization key
    start               start the VM Analytics collector
    stop                stop the VM Analytics collector
    restart             restart the VM Analytics collector
    config              import and list VM Analytics collector config
    list                list configured vCenter instances
    connect             connect to a new vCenter instance
    disconnect          disconnect from a vCenter instance
    enable              enable stat collection for a vCenter instance
    disable             disable stat collection for a vCenter instance
    version             view the current VM Analytics collector version
    phonehome           phonehome the logs now
    stackdump           dump the stackframes to vm_analytics.log
    test                diagnose VM Analytics collector
 
optional arguments:
  -h, --help            show this help message and exit
  -v, --verbose         enable verbose output

Common Commands

Register the Collector
Import the Configuration of the Previous OVA Collector
Verify Connectivity to Pure1 
Display the Collector Version
Generate a Support Bundle
List vCenter Connections
Connect a vCenter
Remove a vCenter Connection 
Change pureuser's Password 
Upgrade a v3.0 Collector
Collector CLI Commands (summary table) 

Register the Collector

Register the collector with Pure1 using the purevmanalytics register command and the authorization key from the Pure1 Manage the VM Topology > Collector Configuration page. Registration is required only once, at the first installation of v3.0.0 collector.

$ purevmanalytics register abcd0123-abcd-0123-abcd-0123abcd0123
Registration completed successfully.
Import the Configuration of the Previous OVA Collector

The purevmanalytics config import command gives you the ability to import a configuration (vCenter connection info) so you don’t need to re-connect your vCenters that you’ve connected in the past.

The import functionality does not support the following import scenarios:

  • To to from an on-array collector
  • From a v3 collector to another v3 collector
  • From a v2 collector to another v2 collector.
purevmanalytics config import --host HOST [--user USER]

Please note that HOST refers to the hostname of the previous OVA. The purevmanalytics config import command imports the vCenter connection info from the previous OVA to the newly deployed OVA.


Verify Connectivity to Pure1 

Use the purevmanalytics test pinghome to test the connectivity from the OVA to Pure1 Manage.

$ purevmanalytics test pinghome
Connection to the cloud was successful!
Display the Collector Version
$ purevmanalytics version
Pure Storage VM Analytics Collector Configuration
Version: 3.0.0
Build: 3.0.0-20200925-3e0a406

The numbers shown above are examples. Your numbers may be different.

Generate a Support Bundle

To prepare logs to send to Pure Technical Services or to view log contents:

$ puresupport bundle
Generating bundle...
Support bundle archive available at: https://10.15.87.15/puresupport-bundle-202010220428.tar.gz
Connect with user "pureuser" and temporary password "3f8WMFRxVc9V5jhp"
The support bundle will be available for 1 hour only.

To access the support bundle:

  1. Enter the bundle archive (https://10.15.87.15/puresupport-bund...0220428.tar.gz in the example above) in a browser URL field and hit enter.
  2. Log in as pureuser with the provided temporary password.
        bundle_login.png
  3. The support bundle is downloaded to your downloads folder by default. 
List vCenter Connections

purevmanalytics list

$ purevmanalytics list
password for pureuser: 
Name                      Enabled    State      Duration (s)  Last collection
------------------------  ---------  -------  --------------  -------------------------
example01.purestorage.int  True       IDLE            35.2798  2018-12-21 03:35:47+08:00
Connect a vCenter

purevmanalytics connect --hostname <VCENTER IP ADDRESS> --username <USERNAME@domain.name>

$ purevmanalytics connect --hostname example02.purestorage.int --username user@domain.com

$ purevmanalytics list
password for pureuser: 
Name                      Enabled    State      Duration (s)  Last collection
------------------------  ---------  -------  --------------  -------------------------
example02.purestorage.int  True       IDLE            60.8616  2018-12-21 03:36:13+08:00
example01.purestorage.int  True       IDLE            35.2798  2018-12-21 03:35:47+08:00

 If one connection has been made to a vCenter, another connection can not be added until the first one is removed.

Remove a vCenter Connection Edit section

purevmanalytics disconnect <HOSTNAME>

 $ purevmanalytics disconnect example02.purestorage.int

 $ purevmanalytics list
 password for pureuser: 
Name                      Enabled    State      Duration (s)  Last collection
------------------------  ---------  -------  --------------  -------------------------
example01.purestorage.int  True       IDLE            35.2798  2018-12-21 03:35:47+08:00
Change pureuser's Password 

The collector's default username is pureuserand the default password is also pureuser. The default password is required to be changed prior to using the collector. Upon your first login (via SSH or console), you are required to change the default user password and will be prompted for the new password. For this change, you do not run the pureadmin command.

To update the password for the pureuser account (other than at the first login), please use the pureadmin CLI command:

pureadmin setattr --password

You are prompted first for the current password, then for the new password. The password change is immediate.

Note: Be sure to note the new password. If pureuser cannot log in, you will have to redeploy the OVA to gain access.

Upgrading to the v3.0 Collector

To upgrade the collector, you would simply deploy a new OVA with the latest collector version. Once the new collector is deployed, you can use the import command to import the configuration from the old OVA collector. So the process would consist of:

  1. Deploy the new OVA using the steps above
  2. On the new OVA, run: purevmanalytics config import --host HOST [--user USER].
    1. HOST refers to the hostname of the previous OVA 
    2. This config import command imports the vCenter connection info from the previous OVA to the newly deployed OVA
  3. Delete the old OVA from vCenter and in Pure1

Once you've upgraded to the v3.0 collector, in order to make sure everything runs smoothly please delete any old versions of the collector. 

Collector CLI Commands

While purevmanalytics command is the main collector CLI command, these commands are also supported:

Command Description
pureadmin Allows changing the pureuser password
pureboot Reboots the collector
pureproxy Configures and manages servers to be used as HTTP or HTTPS proxies
puresetup

Completes the collector deployment

(only available when the collector deployment is incomplete)

puresupport Generates a bundle of collector logs to send to Pure Technical Services for debugging
puresw list Shows OVA or collector packages available for upgrade
puresw upgrade Upgrades OVA or collector packages
purevmanalytics

Configures and manages the collector

(only available when the collector deployment is complete)

exit,
logout
Closes the collector CLI shell and the SSH session

 

Pure1 GUI

Selecting a Cluster

To get started, you select a vCenter, datacenter, and cluster in the top left.

Topology Panel

Once you’ve selected your cluster, the area below should populate. Starting with the right panel, we have a SanKey diagram that shows you the topology.

This is the main area that you will use to navigate through the environment to understand where an issue might be occurring. You can click an object (anywhere but the checkbox) to highlight the object and everything that it’s connected to. You can also select multiple objects (for example, multiple VMs). When you highlight, a bar appears across the bottom that shows aggregated metrics for all the highlighted objects in the column. When you hover over an object in the right panel, the tooltip displays at the top of the left panel in order to not cover up any of the topology. 

The tooltip displays read and write latency, IOPS, and bandwidth metrics for the object. For VMs and datastores, capacity metrics are optionally displayed.

You can sort by any of the columns, by clicking on the column header, and you can filter by typing in the search box. The lines connecting the objects are representative of the relative bandwidth between the objects - the thicker the line, the larger the bandwidth.

Modifying the Topology

In the top right corner, you have a couple of ways to change what you see in the topology.

First, the dropdown in the middle allows you to choose which metric to show on the objects below. This also determines the sort order. The Average and Max is calculated over the time period selected on the right. Pure1 stores up to 7 days worth of metrics on the environment.

The Show Historical checkbox allows you to see any objects that may no longer exist in the most recent portion of the time window. By default, we only show the latest topology in the selected time range in order to simplify the navigation, but in a virtualized environment, things change - and this feature allows you to follow changes back in time.

To the right of Show Deleted, you can click the filter icon to remove everything that’s not highlighted from the Topology. This allows you to drill down into just the information that you want to see. Clear brings the Topology back to the original state, but won’t reset the time window or metrics.

Capacity Metrics

Capacity metrics are available only for vSphere instances that are connected with a v.3.0.0 or higher collector.

Capacity Metric Description
Capacity - Used Raw capacity used in the VM or datastore
Capacity - Used (%)

Percentage of the VM's or datastore's partitioned capacity that is used

(Raw capacity as a percentage of total partitioned capacity on the VM or datastore)

Capacity - Churn

A measure of how much capacity has changed within the selected time period. Churn is based on the following metrics:

  • The maximum used capacity over the selected time period
  • The minimum used capacity over the selected time period
  • The total provisioned capacity

Capacity Churn

Capacity Churn is a new metric that we introduced in order to help quickly identify VMs or Datastores that have had a large change of capacity in the selected time period. Capacity Churn does not take into account the direction of the change, so a VM or Datastore that's decreased by a large amount of capacity and a VM that's gained a large amount of capacity can be analyzed side-by-side. 

To display a capacity metric in the VM Topology page, select it in the metrics drop-down menu:

capacity_menu.png

See the Performance Page for descriptions of other metrics.

Graphs

In order to graph the metrics for an object, click the checkbox in the top right of the object.

Here, we’ve selected the graphs for the Volume. You can drag the slider in between the two panels to make the graphs bigger, and you can click-and-drag on the graph to zoom into a particular time range.

When a capacity metric is selected, a capacity metrics box appears in the graphs panel with graph lines for selected VMs and datastores:

capa_leftbox.png

Click partitions to expand the graphs:

capa_graphs.png

Collected Metrics

Storage Adapter Metrics

Metrics collected using the HostSystem managed object.

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/storage_adapter_counters.html

'storageAdapter.numberReadAveraged.average'
'storageAdapter.numberWriteAveraged.average'
'storageAdapter.totalReadLatency.average'
'storageAdapter.totalWriteLatency.average'
'storageAdapter.read.average'
'storageAdapter.write.average'

Disk I/O Counters

Metrics collected using the HostSystem managed object.

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/disk_counters.html

'disk.read.average'
'disk.write.average'
'disk.deviceReadLatency.average'
'disk.kernelReadLatency.average'
'disk.totalReadLatency.average'
'disk.deviceWriteLatency.average'
'disk.kernelWriteLatency.average'
'disk.totalWriteLatency.average'
'disk.queueLatency.average'
'disk.queueReadLatency.average'
'disk.queueWriteLatency.average'
'disk.totalLatency.average'
'disk.totalReadLatency.average'
'disk.totalWriteLatency.average'

Virtual Disk Counters

Metrics collected using the VirtualMachine managed object.

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/virtual_disk_counters.html

'virtualDisk.numberReadAveraged.average'
'virtualDisk.numberWriteAveraged.average'
'virtualDisk.totalReadLatency.average'
'virtualDisk.totalWriteLatency.average'
'virtualDisk.read.average'
'virtualDisk.write.average'

Datastore Counters

Metrics can be collected using the HostSystem and/or VirtualMachine managed object, but we'll only do it against the Host, as it's the same datastores we're inquiring about.

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/datastore_counters.html

'datastore.datastoreMaxQueueDepth.latest'
    (This is equivalent to DQLEN on esxtop)
'datastore.numberReadAveraged.average'
'datastore.numberWriteAveraged.average'
'datastore.totalReadLatency.average'
'datastore.totalWriteLatency.average'
'datastore.read.average'
'datastore.write.average'

CPU and Memory Counters

These are collected on the HostSystem and VirtualMachine objects.

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/cpu_counters.html

https://www.vmware.com/support/developer/converter-sdk/conv61_apireference/memory_counters.html

'cpu.usage.average'
'mem.usage.average'

Information Topology

The topology of the information below reflects the topology used within vSphere.

vSphere

name:str 
datacenters:[datacenter]

Datacenter

name:str
datastores:[datastore]
vms:[vm]
hosts:[host]
clusters:[cluster]

Storage Pod

name: str

vApps

name: str

VM Folders

name: str

Cluster

name:str
hosts:[host name strings]

Datastore

name:str
mounts:['host1', 'host2', ...]
vms:['vm1', 'vm2', ...]
vmfs_version:str
capacity:long
type:str 'VVOL', 'VMFS'
accessible:boolean
free_space:long

Host

name:str
devices:[device] – storage devices
adapters:[adapter]
power_state:str – e.g. poweredOn or poweredOff, or standBy
standby:str
esxi_version:{build:str, full_name:str, version: str}

Device

canonical_name:str - 'naa.624a937072105f4fe795424400011010'
name:str - '/vmfs/devices/disks/naa.624a937072105f4fe795424400011010'
path:str - '/vmfs/devices/disks/naa.624a937072105f4fe795424400011010'
displayName:str - 'PURE iSCSI Disk (naa.624a937072105f4fe795424400011010)'
queueDepth: int
ssd: boolean
vendor: str - 'PURE'

Adapter

device:str - 'vmhba65'
model:str - 'iSCSI Software Adapter'
name:str - 'iqn.1998-01.com.vmware:3mt-vm90-5f940c1e'
iscsi_name:str

VM

name:str
host:str - 'host1'
disks:[vdisk]
guest_os:str
guest_id:str
latency_sensitivity:str - 'Normal', '?'
datastore:str
power_state:str
devices:
partitions:[partition]

Virtual Disk

file:str - '[VVol Datastore] rfc4122.773f0d5c-f442-49ea-b7f1-9de26e972e5c/test-dan-00.vmdk'
datastore:str
label:str
instance_id:str (this is the disk controller ID) – e.g. scsi0:1
naa: str – For RDM disks only, to map to the volume

Partition

capacity:long
disk_path:str
filesystem_type:str
free_space:long
mappings:[vdisks]


Useful Links

This tool is the first step in the troubleshooting process, but that process will often take you outside of Pure Storage. Here are some useful links where you might be able to find solutions:

https://kb.vmware.com/s/article/2001003 - VMware's KB article on troubleshooting VM performance issues. 

https://www.codyhosterman.com/ - Cody Hosterman's blog. This blog is a never-ending source of great information about anything and everything VMware.

https://support.purestorage.com/Solu...Platform_Guide - Everything you need to know about anything Pure Storage and VMware, including vSphere (ESXi/vCenter), vRealize, Virtual Volumes, Site Recovery Manager, backup integration, and other ecosystem solutions.

Important Caveats

vCenter Metrics

The collector pulls metrics via API from vCenter - this means that any caveats from vCenter show up here as well. Some points that we're aware of are as follows:

The minimum granularity for latency is 1ms. This means that if a VM is running at 0.56ms latency (via esxtop), it shows up as 0ms in vCenter, and by extension, in our VM Analytics product. VMs that are consistently running at sub-ms latency will show up in Pure1 as "<1ms" in order to prevent misrepresentation. 

"Unknown" Arrays

We pull metrics from vCenter on disks, VMs, hosts, and datastores, regardless of what array the VMs are on. This means that the VMs that are not on Pure arrays show up as "Unknown" arrays in Pure1. Since we do not pull metrics from the arrays themselves, any non-Pure arrays or volumes do not have metrics associated with them.