/ 
Sys-Monitor-Lite-0.07
/ README.md

Sys::Monitor::Lite

A lightweight system monitoring toolkit. Using the script/sys-monitor-lite script you can collect CPU, memory, disk capacity, disk I/O, network, and other Linux metrics in JSON or YAML format. It runs entirely on Perl with no external dependencies.

Features

Installation

To install from CPAN:

cpanm Sys::Monitor::Lite

To use it directly from the repository:

git clone https://github.com/yourname/sys-monitor-lite.git
cd sys-monitor-lite
perl Makefile.PL && make install

You can also run the scripts in the repository directly without installing.

Usage (Command Line)

Collect metrics once

script/sys-monitor-lite --once

Collect continuously at 5-second intervals (default)

script/sys-monitor-lite --interval 5

Limit the metrics collected and output as JSON Lines

script/sys-monitor-lite --interval 10 --collect cpu,mem,disk --output jsonl

Output as YAML

script/sys-monitor-lite --once --output yaml

Prometheus text output (node_exporter alternative)

script/sys-monitor-lite --once --output prometheus --prefix sysmon_

Add --labels host=... to attach fixed labels and --timestamp to append sample timestamps to each metric line.

Key options

| Option | Description | | ----------- | ---- | | --interval <seconds> | Interval for repeated collection. Defaults to 5 seconds. Values ≤ 0 collect only once. | | --once | Collects metrics once. Equivalent to omitting --interval. | | --collect <list> | Comma-separated list selecting from system,cpu,load,mem,disk,disk_io,net,process. | | --output <format> | Choose json (default), jsonl, yaml, or prometheus. | | --pretty | Format JSON output (jsonl ignores this). | | --prefix <name> | Prefix metric names (useful for --output prometheus, e.g. sysmon_). | | --labels <k=v,...> | Attach fixed labels to Prometheus output (e.g. --labels host=web1,role=app). | | --timestamp | Append timestamps to all outputs (Prometheus uses the sample timestamp at the end of each line). | | --check | Run once, evaluate thresholds, and exit with Nagios-style status codes (0=OK, 1=WARN, 2=CRIT). | | --warn <expr> | Threshold expression like mem.used_pct>80. Can be repeated. Implies --check. | | --crit <expr> | Critical threshold expression like mem.used_pct>90. Can be repeated. Implies --check. | | --top <field=count> | Collect the top N processes by cpu or rss (e.g. --top cpu=5). Automatically enables the process metric. | | --watch <names> | Comma-separated list of process names/commands to include (e.g. --watch nginx,sssd). | | --help | Show help (POD). |

The JSON output can be combined with tools such as jq or jq-lite, and YAML output works nicely with tools like yq.

script/sys-monitor-lite --once | jq '.mem.used_pct'

Threshold-based checks

For cron, Ansible, or Nagios-style alerting you can evaluate thresholds and exit with status codes:

script/sys-monitor-lite --check \
  --warn mem.used_pct>80 \
  --crit mem.used_pct>90

The command prints a compact summary such as OK - mem.used_pct=42.1 (>80 >90) and exits with 0/1/2 for OK/WARN/CRIT, respectively.

Process monitoring shortcuts

You can request lightweight process data without pulling in top or ps:

# Top 5 CPU consumers
script/sys-monitor-lite --once --top cpu=5

# Top 5 memory consumers and specific daemons
script/sys-monitor-lite --once --top rss=5 --watch nginx,sssd

The --top and --watch switches automatically enable the process metric, returning PID, command, CPU %, RSS, state, threads, and UID for the matching processes.

Using as a Perl Module

use Sys::Monitor::Lite qw(collect_all to_json);

my $metrics = collect_all();
print to_json($metrics, pretty => 1);

If you prefer YAML, call to_yaml instead:

print Sys::Monitor::Lite::to_yaml($metrics);

Instead of collect_all, you can pass an array reference like collect(["cpu", "mem"]) to specify which metrics to gather.

Available Data

License

MIT License

Author

Shingo Kawamura (@kawamurashingo)