Key Linux Server Performance Metrics, Monitoring Tools, and a Python Script for Automated Data Collection

Performance testing of Linux servers requires attention to a variety of critical metrics to ensure efficient and stable operation. The most important indicators include:

1. CPU usage : user time, system time, idle time, and I/O wait time.

2. Memory usage : total, used, available memory, cache, buffers, swap space, and swap usage.

3. Disk I/O : read rate, write rate, IOPS, and average queue length.

4. Network bandwidth : send/receive rates, network errors, and packet loss.

5. Process information : number of processes, zombie processes, and load average.

6. File system : mount points, usage percentage, and free space.

7. System logs : files under /var/log that record events and errors.

8. Boot and response times : time taken to start the system and to respond to requests.

9. Context switches : number of switches per second.

10. Interrupts : hardware interrupt count per second.

To collect these data, a range of command‑line tools can be used, such as top , htop , vmstat , iostat , mpstat , dstat , sar , nmon , netstat , ss , iptraf , iftop , tcpdump , Wireshark , iotop , sysdig , strace , and ltrace . Example commands include:

top -b -n 1 | grep "Cpu(s)"

free -h

iostat -x 1 1

ifstat 1 1

ps aux --sort=-%cpu

df -h

tail -f /var/log/syslog

vmstat 1 1

A practical Python script using psutil and openpyxl automates this monitoring: it records CPU, memory, and disk usage every second, prints the values, and writes a batch of 60 samples (one minute) to an Excel file, also calculating average metrics.

The script initializes an Excel workbook, defines get_system_info() to fetch the three metrics, and record_data_to_excel() to append data and averages to the sheet. A continuous loop gathers timestamps, calls the fetch function, prints the snapshot, stores the data, and sleeps for one second. When a minute’s worth of data is collected, it is saved to system_monitor.xlsx . Interrupting the script with Ctrl+C triggers a final save of any remaining data.

Running the script with python3 system_monitor.py on a Linux terminal starts real‑time monitoring and produces an Excel file containing per‑second measurements and per‑minute averages, providing a convenient record for performance analysis.