Evaluating the Impact of Self-Healing Mechanisms in Monolithic and Microservices Architectures

Microservices are powerful but fragile if not kept under close watch. Ensuring stability requires more than collecting metrics; it demands rapid fault recognition and reliable recovery when failures occur. Legacy systems, which relied on fixed thresholds and reactive alerts, are no longer sufficient. They struggle with unpredictable workloads and chaotic service-to-service communication typical of distributed systems. The result is familiar: problems are detected too late, recovery is prolonged, and operational burdens increase. This work presents a self-monitoring microservices dashboard designed to expand observability. The dashboard can flag anomalies, predict escalations, and trigger corrective actions. At its core is an integrated module that replaces static rules with adaptive intelligence. Built to integrate open-source tools such as Prometheus and Grafana, the framework automates monitoring and response. Static thresholds are replaced with unsupervised learning to detect abnormal patterns and reinforcement learning to determine corrective actions. As a result, the dashboard learns and self-regulates, enabling proactive observability and resilience in microservices. This approach reduces downtime, accelerates recovery, and lessens the operational load, offering a practical step toward autonomous system management.