Skip to content

Infrastructure Discovery

Purpose

Infrastructure Discovery is a tab-based Clarity Transform dashboard of the infrastructure estate: compute, devices, databases and rightsizing. Use it during discovery to understand the environment at a glance before planning move groups and migration waves. The on-screen subtitle reads: "Comprehensive view of your infrastructure assets including compute, devices, and databases across all datacenters."

The KPIs and charts are built from the records in the CMDB, which the integrations and CSV Import populate. The dashboard is read-only, apart from the analyst notes on the Rightsizing tab: saving a note changes data.

Where to find it

Discover → Data Analysis → Infrastructure Discovery. The page header is Infrastructure Discovery Dashboard.

What you see

The dashboard has five tabs: Executive Summary, Compute Infrastructure, Devices, Databases and Rightsizing. The Executive Summary tab loads when the page opens. Each other tab loads the first time you open it, and keeps its figures until you reload the page. If a tab's figures cannot be loaded, the page shows no error message: the cards do not update and the charts stay blank or keep their loading spinner. Reload the page, and if that does not help, report it.

Hover over a chart for its values. Each chart has icons at its top right: Restore, which resets the chart's view (it does not reload the data), and Save as Image, which downloads the chart as a picture. Bar charts also have Zoom and Zoom Reset icons. A chart with nothing to show displays a message instead, for example "No OS data available" or "No cluster data available".

Executive Summary

Cards:

  • Datacenters
  • Clusters
  • Hosts
  • Devices
  • Databases

Charts:

  • Asset Distribution: a doughnut of the record counts by asset type.
  • In-Scope vs Out-of-Scope (Devices): a doughnut with the segments In Scope, Out of Scope and Unknown.
  • Discovery Health (Recency): one stacked bar with the segments Recent <1d, Last 7d, Older >7d and Unknown.
  • Stale Inventory (Last Seen): the records grouped by the age of their last-seen date: 0-7d, 8-30d, 31-90d, 90+d and Unknown.
  • Discovered Services (Listening Ports): the number of hosts running each service found by an open-port scan. Until a scan has found any, the chart reads "No discovered listening-port services yet. Run a network scan."

A large Older >7d, 90+d or Unknown share means discovery has not run recently, or that records are no longer returned by their source. Check the source on Discovery Sources.

Compute Infrastructure

Host cards:

  • Total Clusters
  • Total Hosts
  • Total Sockets
  • Total Cores
  • Total RAM (Hosts), in TB.

VM cards:

  • Total VMs
  • Total vCPUs (VMs)
  • Total vRAM (VMs), in TB.
  • Total Storage (VMs), in TB.

Overcommit cards. Overcommit is how much virtual capacity is allocated for each unit of physical capacity. Each card shows a ratio such as "2.4 : 1" and a rating:

  • CPU Overcommit: "vCPU : pCPU".
  • RAM Overcommit: "vRAM : pRAM".
  • Storage Overcommit: "Provisioned : Capacity".
Rating Ratio
OK 3 : 1 or less
Moderate above 3 : 1, up to 5 : 1
High above 5 : 1

A card reads "0.0 : 1" and OK when the ratio could not be calculated. Check the host and VM cards before you rely on an OK rating.

Charts:

  • VMs by Managed By: the VMs per management server, for example per vCenter.
  • VMs by Datacenter
  • VMs by Cluster

Devices

Cards:

  • Total Devices
  • In Scope
  • Powered On
  • Powered Off

Charts:

  • Devices by Technology
  • Operating System Distribution
  • Devices by Region (Top 10)
  • Devices by Business Unit (Top 10)
  • Devices by Environment

Databases

Cards:

  • Total Databases
  • Total Size, in GB.
  • Unique Hosts
  • Clustered

Charts:

  • Databases by Type
  • Databases by Environment
  • Databases by Host (Top 15)

Rightsizing

Rightsizing compares the CPU and memory each VM is given with what it actually uses, so you can see which VMs could be made smaller in the target. The tab is filled from VM assessment data collected by PowerCLI discovery (PowerCLI is VMware's command-line toolkit). Until that data has been collected, the VM Utilisation Detail grid reads "No assessment data" and tells you to run PowerCLI discovery.

The figures use P95 utilisation: the level the VM stays at or below for 95% of the observed time. It ignores short peaks, so it is a fairer basis for sizing than the average or the peak.

Cards:

  • Assessed VMs: the VMs with assessment data.
  • CPU Under-utilised
  • Mem Under-utilised
  • Dual Over-provisioned: VMs that are under-utilised on both CPU and memory.
  • Zombie VMs: VMs that look idle, and are candidates for decommissioning rather than migrating.
  • Avg CPU P95
  • Avg Mem P95
  • vCenters

Charts:

  • CPU P95 Distribution: the VMs grouped by CPU P95 range.
  • Memory P95 Distribution: the VMs grouped by memory P95 range.
  • VMs by Managed By: the assessed VMs per vCenter. If it reads "No vCenter data available", the VMs' Device records have no Managed By value.
  • Top Over-provisioned VMs (Lowest CPU P95): the VMs with the lowest CPU P95. Hover over a bar for the VM's CPU P95, vCPUs and memory P95.

The VM Utilisation Detail grid has one row per assessed VM, sorted by VM Name until you change the sort, and the columns:

  • VM Name: a red Zombie label follows the name of a VM flagged as a zombie.
  • vCenter
  • vCPUs (Cur→Rec): the current and the recommended number of vCPUs.
  • RAM GB (Cur→Rec): the current and the recommended RAM.
  • CPU Avg%
  • CPU P95%: shown in red below 10%, and in amber from 10% up to 20%. These are the strongest candidates for fewer vCPUs.
  • CPU Peak%
  • Mem Avg%
  • Mem P95%: shown in red above 85%, where the VM may need more memory, and in amber below 20%, where it may need less.
  • Mem Peak%
  • Disk IOPS
  • Net In/Out
  • Obs Days: the number of days of data behind the figures.
  • Confidence: High (green), Medium (amber) or Low (red).
  • Notes: generated text, for example "CPU P95 critically low (x%) - strong vCPU reduction candid…", "Zombie VM candidate - CPU P95 x%, Mem P95 y%. Consider d…" or "Short observation window (0d) - extend data collection…". Notes longer than 60 characters are cut short; hover over the note for the full text. Select the note to open the Analyst Note dialog. A VM with no note shows an Add note button instead. A VM with no provider id (the identifier its source platform gives it) shows its note as plain grey text, and you cannot add or change a note for it.

A dash in a figure column means no value was recorded.

The grid toolbar has a back arrow, which returns to the previous page in your browser, and Export CSV, which downloads the visible columns. The rows have no row actions.

Analyst Note dialog

Selecting a VM's note, or its Add note button, opens the Analyst Note dialog. It shows the VM name and one unlabelled text box, filled with the current note (placeholder "Enter analyst note…"). Help: "Saved notes override the auto-generated observation." The buttons are Cancel and Save Note.

Save Note reads "Saving…" while the note is saved. When it succeeds, the dialog closes and the row shows the new note. The note is saved in Clarity Transform against the VM, so other users see it too. Saving an empty box clears the note, and the row then shows Add note.

If the save fails, a browser message reads "Failed to save: " followed by the reason, or "Failed to save note" and asks you to try again. The dialog stays open, so you can retry or select Cancel.

Filters

The VM Utilisation Detail grid has a Search: box and a filter row. It remembers its Search: text, column filters, sort order and page size in your browser between visits. If the grid shows fewer rows than you expect, clear the Search: box and the filter row.

How to

Size the estate

  1. Go to Discover → Data Analysis → Infrastructure Discovery. The Executive Summary tab opens.
  2. Read the Datacenters, Clusters, Hosts, Devices and Databases cards.
  3. Read Discovery Health (Recency) and Stale Inventory (Last Seen). Large Older >7d, 90+d or Unknown values mean the figures may be out of date.
  4. Select Compute Infrastructure and read the cards for hosts, VMs and overcommit ratios.

Check device scope

  1. On the dashboard, select the Devices tab.
  2. Compare In Scope with Total Devices.
  3. To change a device's scope, edit its Scope on the Device grid, then reload the dashboard.

Find rightsizing candidates

  1. On the dashboard, select the Rightsizing tab. If VM Utilisation Detail reads "No assessment data", run PowerCLI discovery first.
  2. Read the CPU Under-utilised, Mem Under-utilised, Dual Over-provisioned and Zombie VMs cards.
  3. In VM Utilisation Detail, compare vCPUs (Cur→Rec) and RAM GB (Cur→Rec) for each VM, and read its Notes. Check Obs Days and Confidence before you rely on a recommendation. A generated note that starts "Short observation window" asks you to collect more data.
  4. To export the grid, select Export CSV.

Add an analyst note to a VM

  1. On the Rightsizing tab, in VM Utilisation Detail, select the VM's note in Notes, or Add note if it has none. The Analyst Note dialog opens. A VM with no provider id cannot be given a note.
  2. Type the note in the text box. To remove a note, clear the box.
  3. Select Save Note, or Cancel to close the dialog without saving. A saved note overrides the generated observation.