Terminal UI

Run vsfleet to browse inventory, switch vCenters, and inspect resources. Press ? on any screen for its keys.

Browse screen

Key	Action
`1`–`7`	VMs, templates, hosts, clusters, datastores, networks, vApps
`h` / `l`, `←` / `→`	Previous/next resource tab
`k` / `j`, `↑` / `↓`	Previous/next row
`Enter`	Open selected resource
`c`	Manage contexts
`a`	Toggle current/all contexts
`/`	Filter table
`Tab`	Widen filter to estate search
`Esc`	Clear filter, restore view, or go up a folder
`s` / `S`	Open the sort menu / reverse the sort
`r` / `R`	Reload current/all contexts
`d`	Diagnose selected row's vCenter
`space` / `t` on Networks	Fold switch / toggle tree and plain list
`t` on VMs or Templates	Toggle folders and plain list
`v` on Networks	Open VLAN map
`H`	Open History
`?` / `q`	Help / quit

The selected row is marked ▸, so it stays visible under NO_COLOR.

Other screens have their own pages: VM dashboard, detail actions and SSH, datastore browser, and networks and clusters.

Filtering and searching

Press / to filter the current view by name. The query line also counts matches elsewhere: /ubuntu 0 VMs in prod · 2 in the estate — tab to widen. Tab searches every kind and vCenter in cached inventory; Tab again or Esc narrows back. ↑/↓ move the selection while typing, and Enter keeps it.

Sorting

Press s to choose the order of the current tab. The menu lists the tab's columns with a letter each: s then m sorts VMs by memory. Choosing the current order again, or r in the menu, reverses it; S reverses it without opening the menu.

Tab	Sorts by
VMs	name, power, CPU, memory, IP address, host, datacenter
Templates	name, guest OS, CPU, memory, disk, datacenter
Hosts	name, state, cluster, CPU used, memory used, VMs, version
Clusters	name, status, hosts, cores, memory, DRS, HA, datacenter
Datastores	name, status, type, capacity, free, used %, datacenter
Networks, vApps	name, status

Numbers start largest first and text A to Z. Values not reported sort last either way, and ties fall back to name. Names and versions compare their numbers, so web-2 comes before web-10. The header shows the order, such as sort: memory ↓, and the sorted column's heading carries the arrow. Each tab keeps its own order, and vsfleet remembers it for the next run. Search results sort by name or status (s toggles).

Browsing by folder

Press t on the VMs or Templates tab to browse folders one level at a time, as in the vSphere client. Enter opens a folder, Esc goes up, and the message line shows where you are: in Taipei › Prod › Web. Folder rows show how many VMs they hold and their vCPU and memory totals. A filter searches the open folder and everything below it. vApps appear in their folder marked ▣; Enter opens the vApp page.

Contexts and vApps

On Contexts (c): Enter selects a context, a selects all, n/e/x add, edit or remove, o logs out, and d diagnoses. Logging out keeps cached inventory and stops refreshing that context.

Browse-screen a shows only vCenters that are loaded. The others are listed below the table with their state:

○ not connected: edge-vc, dr-vc · R connects all · c then enter connects one, d diagnoses
◐ credentials required: site-3 · R asks for them · c then enter picks one
✕ old-vc: connection failed · c select, d diagnose

In vsfleet demo, n opens a read-only preview of the setup form.

A password source that is gone

env:, file: and exec: passwords are read on every connection, so a missing source shows its cause:

✕ lab-vc: LAB_VC_PW is not set · d diagnose
✕ lab-vc: password helper vsfleet-credential failed: Vault is sealed · d diagnose

d explains how to recover: export the variable and restart vsfleet, restore the file and press r, or fix the helper and press r. Press e there to switch to another source; the form tests it before saving. See credentials.

vApp workspace

Enter on a vApp shows its CPU and memory allocation, start order, and nested members. A vApp's limit is shared by its members and can throttle VMs that have no limit of their own. Enter on a member opens it; Esc goes back.

Live charts sum member CPU and active memory against the vApp's limit, or its members' total when unlimited; ▲ marks samples near the limit. Members show peak CPU, CPU ready and CPU limited time, and s sorts by start order or busiest. </> changes the range. Members that cannot be read are named and left out of the totals.

History workspace

Press H for four panes, switched with Tab: Changes, Trends, Runs, and Health. n takes a capture of the current scope in the background.

Changes compares two captures, newest two by default. Migration blockers come first (vanished VMs or hosts, old snapshots, changed firmware or devices), then sizing, growth and churn.
Trends leads with datastores projected to fill within 30 days and growing stale snapshots, then CPU, memory, storage, and VM counts across up to 30 runs.
Runs lists captures with their coverage, duration, label, pin and note.
Health shows the latest migration verdict and findings, worst first. Enter opens a finding, / filters. Use vsfleet analyze health to tune thresholds.

Keys	Action
`b` / `t`	Choose which end of the comparison to move: baseline or target
`←` / `→`	Move that end one run
`R`	Pick a run from a list
`c`	Set the baseline to the newest older run covering the same vCenters
`1`–`4`, `0`	Filter blocks / sizing / growth / churn; clear
`h`	Open the VM timeline from a change
`e` / `N` / `p` in Runs	Edit label / note / pin
`Enter` in Runs	Compare with the run before

The coverage line marks each vCenter ● reached, ✕ not compared, or · no record; c limits the comparison to what both runs covered. vsfleet demo stores nothing, so it refuses n, e, N and p. See assessments for captures and comparisons.

VM timeline

Press h on a VM, a vApp member, or a Changes row for that VM's timeline:

Tab	Shows
`1 Changes`	What changed between stored captures, as far back as your history
`2 vCenter events`	Who did what and when, read live, including failed tasks and changes undone before the next capture; as far back as vCenter keeps events
`3 Combined`	Each stored change beside the event that caused it

On Combined, ← marks the event that caused a change, ‹ an event from before vsfleet first saw the VM, and · an event that left no stored change, for example because it failed or was undone.

Events are read only when you open an events tab or press r. The header says LIVE and names the vCenter; a vCenter that could not be read is named, not left out. Task rows show ✓ or ✕ with vCenter's error. a shows unchanged runs or routine events. The same log is available as vsfleet events --for vm/<name>.

Refresh and cache behavior

Context	Background refresh
Selected	Every 20 seconds
Visited before, not selected	Every 200 seconds
Never visited or logged out	None
Waiting for a password	Only when selected or reloaded
Refresh failed	Keeps cached inventory, with a warning

--refresh 5s polls faster and --refresh -1 turns timers off. Slow vCenters are polled less often.

Large estates

Rows arrive in pages while the cursor stays put, and a load fails only when it stops making progress; --timeout bounds connecting and logging in. Browsing skips virtual disks, guest NICs and snapshot trees, which a capture collects.

Welcome animation

On first run and after an upgrade, a short animation plays while the last vCenter loads; any key skips it. Upgrades also show release notes. Set VSFLEET_NO_WELCOME or CI to turn it off.

Upgrade prompt

When a newer release is out, vsfleet asks at launch:

Key	Action
`y`	Upgrade and restart (Homebrew or `go install`, macOS and Linux)
`c`	Copy the upgrade command or release link (other installs)
`n` / `Esc`	Ask again in a week
`s`	Skip this release

VSFLEET_NO_UPDATE_NOTIFIER=1 turns off checks, prompts and notices.