System Diagnostics

Overview

This repository provides tools for calculating CPU usage and other system diagnostics on Linux systems. This tool is built for efficient monitoring and provide some configuration options.

---

Installation Instructions

Option 1: Using the Build Script

1. Clone the repository:

   git clone https://github.com/jaca230/system_diagnostics.git
   cd system_diagnostics

2. Run the build script:

   ./build.sh

Option 2: Manual Build

1. Clone the repository:

   git clone https://github.com/jaca230/system_diagnostics.git
   cd system_diagnostics

2. Create a build directory and navigate to it:

   mkdir build
   cd build

3. Run CMake to configure the build:

   cmake ..

4. Build and install the project:

   make install -j$(nproc)

Executable and Library Outputs

• The executables will be located in the bin/ directory.
• The libraries will be located in the lib/ directory.

---

Usage

To use the system diagnostics tools, navigate to the bin/ directory and run the desired executable. For example:

cd bin
./system_diagnostics --help
./system_diagnostics

---

Configuration

The behavior of the system diagnostics tool can be customized using a configuration file written in JSON format. Below is a description of each configuration section and its options.

Config File Sections

debug

• verbosity:
  • Description: Controls the level of debug output. Higher values produce more detailed logs.
  • Example:
    • 0: Minimal output (default).
    • 1: Includes additional output and warnings.
    • 2: Includes even more output and warnings, etc.

printer

• print_line_number:

  • Description: Whether to include the source code line number in log messages.
  • Example: true (enabled) or false (disabled).

• print_current_time:

  • Description: Whether to include a timestamp in log messages.
  • Example: true (enabled) or false (disabled).

• prefix and suffix:

  • Description: Strings added before and after log messages, respectively.
  • Example: "system_monitor" for prefix.

• info_color, warning_color, error_color:

  • Description: Define the colors for log messages based on their severity.
  • Options: "white", "yellow", "red", "green", "black", "blue", "magenta", or "cyan"

system_info

• update_period_jiffies:

  • Description: Determines how frequently (in jiffies) CPU usage data is polled and recorded.
  • Note: A jiffy is a system-dependent unit of time (often 10 ms). You can find your system’s jiffy rate using getconf CLK_TCK.
  • Example: 20 jiffies (200 ms at 10 ms/jiffy).

• average_period_jiffies:

  • Description: Specifies the time period (in jiffies) over which the average CPU usage is calculated.
  • Example: 100 jiffies (1 second at 10 ms/jiffy).

---

Example Config File

Located in config/config.json:

{
    "debug": {
        "verbosity": 0
    },
    "printer": {
        "print_line_number": true,
        "print_current_time": true,
        "prefix": "system_monitor",
        "suffix": "",
        "info_color": "white",
        "warning_color": "yellow",
        "error_color": "red"
    },
    "system_info": {
        "NOTE": "A jiffy is a unit defined by your system, usually 10 ms. See `getconf CLK_TCK` for the rate in Hz.",
        "update_period_jiffies": 20,
        "average_period_jiffies": 100
    }
}

---

Notes

• Accuracy: The program may produce inaccurate results if:
  1. The CPU usage is polled too frequently.
  2. The total CPU usage is small.

This is sort of an "uncertainty principle." If the CPU usage is too small in the polled timeframe, the act of measuring it changes the result. For example, if there CPU is at 0% usage in total, and we measure that, it takes some CPU usage to measure that. So measuring each core's usage after that may show a non-zero value.

Furthermore there are small innaccuracies due to rounding errors.