Worked on enhancing monitoring reliability and system stability for the kamailio/kamailio repository by addressing a bug in CPU load parsing. Focused on ensuring that only valid numeric values were processed, the solution reduced the risk of incorrect monitoring metrics and false alerts, directly supporting more accurate operational decisions and capacity planning. The approach involved targeted validation around timer intervals and was implemented in C, leveraging skills in system programming and performance optimization. The fix was validated through continuous integration, resulting in improved accuracy of monitoring data without introducing new features, reflecting a depth of attention to operational robustness and code quality.