Monitoring System Performance in CentOS 7

Key Objectives of Performance Monitoring:

• Monitor statistics such as:

  • Memory usage

  • CPU usage

  • Input/Output (I/O) usage

  • Delay times

  • Queuing

/proc File System:

• CentOS 7 uses the /proc file system for monitoring system performance.

• Access via the cat command with various kernel parameters.

• Definition: A pseudo-file system that interfaces directly with the kernel and exists in memory.

• Kernel Parameters Sample:

  • Table 12-2 provides a partial listing of kernel values. For full listings, check proc man page.

Common Kernel System Values:

• Entry Descriptions:

  • iomem: Memory input/output statistics

  • mounts: Mounted file system statistics

  • stat: Process status in readable format, detailing process resource usage.

  • diskstats: Disk statistics providing read/write performance metrics.

  • swaps: Swap file statistics (similar to virtual memory in Windows 10), indicating swap memory usage.

  • loadavg: Average number of processes ready to run over the last 1, 5, and 15 minutes.

Using /proc for System-Related Tasks:

• System-related tasks include:

  • Viewing performance and memory info for optimization purposes.

  • Viewing or modifying runtime parameters, allowing for dynamic resource adjustment.

  • Viewing hardware information for diagnostics.

  • Viewing network parameters to monitor traffic and connectivity.

  • Viewing statistical information to help in resource planning.

• Command Overview:

  • Navigate to /proc directory:
    bash cd /proc

  • List files in /proc:
    bash ls

  • Retrieve performance information:

  • System and file system stats:
    bash cat stat cat mounts

  • Disk stats:
    bash cat diskstats

  • Swap, load averages, and I/O memory usage:
    bash cat swaps cat loadavg cat iomem

Using the top Command:

• Functionality: Displays system statistics in real-time (default 5-second update).

• Basic Usage:

  • Command: top

  • Displays top CPU processes and their resource utilization.

• Key Fields in top Output:

  • uptime: System up time and average loads for 1, 5, and 15 minutes.

  • processes: Total processes running and their resource demands.

  • CPU states: Breakdown of CPU time across modes:

  • User mode

  • System mode

  • Nice tasks

  • iowait (waiting for I/O)

  • Idle time

  • Mem: Total, free, used memory, shared memory, buffers providing a snapshot of memory health.

  • Swap: Total, available, used swap space.

  • PID: Process ID for ongoing tasks.

  • PPID: Parent process ID for hierarchical process management.

  • UID: User ID of the task’s owner.

  • USER: Username of the task's owner, facilitating accountability.

  • SHARE: Shared memory used by the task.

  • To start top with custom iterations:

  • Command: top -n 5 -d 10.00

  • Note: This command uses significant memory; use it judiciously for troubleshooting. Press q to exit and return to command prompt.

Using the free Command:

• Functionality: Displays amounts of free and used memory, including swap, crucial for memory management.

• Output shows in kilobytes by default; use -m to view in megabytes for easier interpretation.

  • Command: free

Using the watch Command:

• Functionality: Repeatedly runs free to show real-time memory statistics and system health.

• Default refresh every two seconds; can control the interval for flexibility.

  • Command: watch free

• Terminate with Ctrl+C. This allows users to monitor changes continuously and adapt to performance needs.

Using the vmstat Command:

• Overview: Reports on process status, memory consumption, paging activity, I/O operations, and CPU usage, providing a comprehensive view of system performance.

• Execution: For average stats since last reboot or specific intervals, aiding in detailed performance analysis:

  • Command syntax: vmstat [time] [iterations]

  • Example: vmstat 5 10 (reporting every 5 seconds for 10 iterations) to track system performance trends over time.

• Key Fields in vmstat Output:

  • procs: Running and sleeping processes for overall system workload.

  • swap: Amount of memory paged for I/O, signifying active memory management.

  • us: Percentage of CPU time used in user space, indicating application performance.

  • sy: Percentage of CPU time used by the kernel, reflecting the system's operational workload.

  • wa: Percentage of CPU time spent in I/O wait, highlighting bottleneck areas.

  • id: Percentage of CPU time spent idle, giving insight into overall system utilization.

Using the uptime Command:

• Functionality: Displays system uptime, CPU average load for 1, 5, and 15 minutes, active user