|
|
@@ -8,6 +8,49 @@ breadcrumbs:
|
|
|
|
|
|
## Commands
|
|
|
|
|
|
+### General Monitoring
|
|
|
+
|
|
|
+- For more specific monitoring, see the other sections.
|
|
|
+- `htop`:
|
|
|
+ - ncurses-based process viewer like `top`, but prettier and more interactive.
|
|
|
+ - Install (APT): `apt install htop`
|
|
|
+ - Usage: `htop` (interactive)
|
|
|
+- `glances`:
|
|
|
+ - Homepage: [Glances](https://nicolargo.github.io/glances/)
|
|
|
+ - Install (PyPI for latest version): `pip3 install glances`
|
|
|
+ - ncurses-based viewer for e.g. basic system info, top-like process info, network traffic and disk traffic.
|
|
|
+ - Usage: `glances` (interactive)
|
|
|
+- `dstat`:
|
|
|
+ - A versatile replacement for vmstat, iostat and ifstat (according to itself).
|
|
|
+ - Prints scrolling output for showing a lot of types of general metrics, one line of columns for each time step.
|
|
|
+ - Usage: `dstat <options> [interval] [count]`
|
|
|
+ - Default interval is 1s, default count is unlimited.
|
|
|
+ - The values shown are the average since the last interval ended.
|
|
|
+ - For intervals over 1s, the last row will update itself each second until the delay has been reached and a new line is created. The values shown are averages since the last final value (when the last line was finalized), so e.g. a 10s interval gives a final line showing a 10s average.
|
|
|
+ - The first line is always a snapshot, i.e. all rate-based metrics are 0.
|
|
|
+ - If any column options are provided, they will replace the default ones and are displayed in the order specified.
|
|
|
+ - Special options:
|
|
|
+ - `-C <>`: Comma-separated list of CPUs/cores to show for, including `total`.
|
|
|
+ - `-D <>`: Same but for disks.
|
|
|
+ - `-N <>`: Same but for NICs.
|
|
|
+ - `-f`: Show stats for all devices (not aggregated).
|
|
|
+ - Useful metrics:
|
|
|
+ - `-t`: Current time.
|
|
|
+ - `-p`: Process stats (by runnable, uninterruptible, new) (number of processes).
|
|
|
+ - `-i`: Specific interrupt stats (by interrupt IDs) (interrupts per second).
|
|
|
+ - `-y`: Total interrupt and context switching stats (by interrupts, context switches) (events per second).
|
|
|
+ - `-l`: Load average stats (1 min, 5 mins, 15 mins) (total system load multiplied by number of cores).
|
|
|
+ - `-c`: CPU stats (by system, user, idle, wait, hardware interrupt, software interrupt) (percentage of total).
|
|
|
+ - `-m`: Memory stats (by used, buffers, cache, free) (bytes).
|
|
|
+ - `-p`: Paging stats (by in, out) (count per second).
|
|
|
+ - `-s`: Swap stats (by used, free) (total).
|
|
|
+ - `-r`: Storage request stats (by read, write) (requests per second).
|
|
|
+ - `-d`: Storage throughput stats (by read, write) (bytes per second).
|
|
|
+ - `-n`: Network throughput stats (by recv, send) (bytes per second).
|
|
|
+ - `--socket`: Network socket stats (by total, tcp, udp, raw, ip-fragments)
|
|
|
+ - Useful plugins (metrics):
|
|
|
+ - `--net-packets`: Network request stats (by recv, send) (packets per second).
|
|
|
+
|
|
|
### File Systems and Logical Volume Managers
|
|
|
|
|
|
- Partition disk: `gdisk <dev>` or `fdisk <dev>`
|
|
|
@@ -148,14 +191,11 @@ breadcrumbs:
|
|
|
- `iostat [-c] [-t] [interval]`
|
|
|
- Monitor processes:
|
|
|
- `ps` (e.g. `ps aux` or `ps ax o uid,user:12,pid,comm`)
|
|
|
-- Monitor a mix of things:
|
|
|
- - `htop`
|
|
|
- - `glances`
|
|
|
- - `ytop`
|
|
|
+- Monitor a mix of things: See the "general monitoring" section.
|
|
|
- Monitor interrupts:
|
|
|
- `irqtop`
|
|
|
- `watch -n0.1 /proc/interrupts`
|
|
|
-- Stress test with stress-mg:
|
|
|
+- Stress test with stress-ng:
|
|
|
- Install (Debian): `apt install stress-ng`
|
|
|
- Stress CPU: `stress-ng -c $(nproc) -t 600`
|
|
|
|
Nordstrand