Self-Healing Workflow Orchestration Framework Using System-Aware Adaptive Recovery for Fault-Tolerant Distributed Systems

This work presents a modular self-healing workflow orchestration framework for improving reliability in distributed systems through automated failure detection and recovery. The framework combines Redis-based task queuing, PostgreSQL-backed persistent workflow state, concurrent worker execution, runtime monitoring, and a pluggable recovery engine that supports multiple recovery strategies. Four recovery policies are evaluated under identical experimental conditions: No Recovery, Fixed Retry, Exponential Backoff, and System-Aware Adaptive Recovery. Each strategy is tested across five independent benchmark runs of 100 workflow instances under controlled failure injection. Performance is evaluated using workflow success rate, workflow failure rate, recovered tasks, recovery latency, throughput, and worker-level execution statistics. The results show that automated recovery substantially improves workflow reliability relative to the No Recovery baseline. Fixed Retry achieved the highest average workflow success rate of 84.4% and the lowest average failure rate of 15.6%, while System-Aware Adaptive Recovery achieved comparable reliability and provides a modular foundation for future runtime-aware and learning-based recovery policies. Status: Preprint / unpublished research manuscript.