vminfo watch

Collects snapshots repeatedly until the requested count is reached or the process is interrupted. It works on Linux, macOS, and Windows and does not require the TUI.

Usage

vminfo watch
vminfo watch --json
vminfo watch --count 1
vminfo watch --interval 2s

Options

Option	Default	Purpose
--json	off	Emit one compact JSON object per line
--interval <duration>	1s	Sampling window for each snapshot
--count <n>	0	Stop after n snapshots; 0 means keep running

--count must be zero or greater and --interval must be positive. Use Ctrl+C to stop an unlimited stream.

Text output

Text mode prints a timestamp followed by resource and connection summaries:

[2026-07-14T12:00:00Z] host=example-host os=linux
cpu=4.2% mem=3.0 GiB/8.0 GiB swap=0 B/2.0 GiB disk=18.0 GiB/80.0 GiB
load=0.18 0.21 0.25 net=↓ 12.0 KiB/s ↑ 3.0 KiB/s conn=tcp 42 udp 8 proc 126 uptime=7d 2h 0m

The next sample is separated by a blank line.

JSON Lines output

JSON mode is suitable for line-oriented pipelines. Each line is a complete object with its own collection timestamp:

{"collected_at":"2026-07-14T12:00:00Z","static":{"os":"linux","hostname":"example-host"},"stats":{"cpu":4.2,"mem_used":3221225472}}

Do not parse JSON Lines as one JSON array. Process each line independently or collect a bounded number of samples first.

Good for

• short troubleshooting windows without opening the TUI
• log and telemetry pipelines that accept JSON Lines
• CI checks that need a fixed number of samples
• comparing a host before and after a workload change

Examples

# Emit exactly three JSON records.
vminfo watch --json --count 3

# Capture one sample in the same envelope used by a stream.
vminfo watch --json --count 1 > sample.jsonl

# Use a slower sampling window for a five-sample observation.
vminfo watch --interval 5s --count 5

The command rejects positional arguments. If collection or output fails, it stops instead of silently skipping a sample.

Related

• summary
• ps
• HTTP API