| AI-Assisted Adaptive Circuit Breaker Monitoring and Failure Prediction for Distributed Cloud Platforms | |
|
MSRDG International Journal of Computer Scientific Technology & Electronics Engineering
© 2026 by MSRDG IJCSTEE Journal Volume 2 Issue 3
Year of Publication: 2026 |
Paper Download Article ID MSRDG-IJCSTEE-V2I3P104 |
|
Abstract: The circuit breaker pattern is a widely used resilience mechanism in distributed cloud platforms, preventing cascading failures by halting requests to unhealthy services once error thresholds are exceeded. However, conventional circuit breakers rely on static, manually configured thresholds that cannot adapt to dynamic traffic patterns, seasonal load variation, or gradually degrading service health, often resulting in delayed reactions or unnecessary service interruptions. This paper proposes an AI-assisted adaptive circuit breaker framework that combines continuous telemetry monitoring with a machine learning-based failure prediction module to dynamically adjust circuit breaker thresholds and state transitions in distributed cloud environments. The framework ingests real-time metrics such as latency, error rate, throughput, and resource utilization, estimates a rolling failure probability using a time-series prediction model, and uses this signal to drive adaptive transitions between the closed, half-open, and open circuit breaker states. [Summarize your key results here once evaluation is complete, e.g., reduction in cascading failures and recovery time compared to static-threshold baselines.] The proposed approach is evaluated against static-threshold and purely reactive circuit breaker baselines using standard reliability and classification metrics, demonstrating the potential of predictive, AI-informed resilience mechanisms for large-scale distributed systems. |
|
| Keywords: Circuit Breaker Pattern; Failure Prediction; Distributed Systems; Cloud Reliability Engineering; Anomaly Detection; Adaptive Thresholding; Microservices | |
