VM dashboard
Press Enter on a VM to inspect its properties and performance. Use the
detail actions to copy fields, connect over SSH, or open the
vSphere Client.
VM detail dashboard
Long property values wrap beneath their labels. The field cursor selects the
full value, including wrapped DNS names and inventory paths; copying preserves
it. Use PgUp/PgDown to scroll tall values.
Health marks use the same default thresholds as History Health:
| Field | ✓ |
▲ |
|---|---|---|
| VMware Tools | running | not installed, not running on a powered-on VM, or needs an upgrade |
| Snapshots | none, or oldest younger than 30 days | oldest 30 days or older, with age shown |
| Guest disks | fullest filesystem has at least 10% free | fullest filesystem has less than 10% free |
Missing evidence gets no mark: an empty Tools state, unread VM configuration, or no filesystems reported by Tools.
Chart pages
Charts sit beside the properties, or below them under 100 columns. An action menu can also move charts below the properties; closing it restores the layout. All five pages share one read, so changing pages makes no vCenter request.
| Key | Page | Shows |
|---|---|---|
0 |
Overview (default) | CPU usage, active memory, disk read + write and peak latency; CPU ready, co-stop, CPU limited, balloon, swap-in, network throughput, dropped packets, sizing signal |
1 |
CPU | Usage, ready, co-stop, CPU limited; “no CPU limit set” when none exists |
2 |
Memory | Active and consumed memory as a share of configured memory, balloon, swap-in rate |
3 |
Disk | Read/write throughput, worst disk latency, IOPS |
4 |
Network | Received/transmitted throughput, dropped packets |
Titles show average, 95th percentile, and peak, subject to the minimum sample
count and window coverage used by performance summaries.
Each chart column keeps its largest sample. A faint baseline means zero, a dot
means no samples, and ▲ marks a reading at or above these thresholds:
| Reading | Threshold | Basis |
|---|---|---|
| CPU usage, active memory | 60% | sizing signal |
| CPU ready | 5% per vCPU | sizing signal |
| Balloon | 1 MiB | sizing signal |
| Co-stop | 3% per vCPU | common VMware guidance |
| CPU limited | 1% per vCPU | measurable throttling |
| Disk latency | 20 ms | common VMware guidance |
| Swap-in rate | 1 KB/s | active swapping |
| Dropped packets | 1 per sample | any drop |
These chart marks are separate from health rules and are neither stored nor exported.
Ranges and refresh
| Key | Action |
|---|---|
0–4 |
Switch chart page |
< / > or , / . |
Shorter / longer range: 1h, 24h, 7d, 30d |
r |
Re-read the range |
← / → |
Change VM, keeping the page and range |
The 1h range uses 20-second realtime samples; 24h, 7d, and 30d use 5-minute, 30-minute, and 2-hour vSphere roll-ups. The window ends show clock times for 1h, dates and times for 24h, and dates for 7d/30d. “As of” is the read time. Roll-ups publish after their interval closes, often leaving a 5 to 30 minute gap at the end.
| Range | Refresh interval |
|---|---|
| 1h | 20 seconds |
| 24h | 5 minutes |
| 7d, 30d | 10 minutes |
Refresh runs while the pane is open. Returning to an expired cached read refreshes it. The previous charts remain during a refresh; failure preserves them with a warning and read time. A negative background refresh interval disables chart refresh too and removes “live” from the header.
Only the open VM makes a read-only QueryPerf request. Device disk/network
counters are summed when no VM total exists. Dashboard counters are separate
from stored assessment performance. Longer ranges depend on the vCenter's
statistics level; unavailable counters show “no samples” and a reason. Where
mem.active is not kept, “Memory active” draws
mem.usage and says so in its title. An
initial failure replaces charts with an error while properties remain usable.
vsfleet demo shows fixed synthetic charts.