Diagnoses and resolves Pegasus workflow failures by analyzing logs, run directories, and common error patterns.
The Pegasus Workflow Debugger is a specialized diagnostic skill designed to streamline the troubleshooting of complex scientific workflows managed by the Pegasus Workflow Management System. It systematically gathers diagnostic information from pegasus-analyzer and job logs, matches failures against a curated database of common patterns—such as file staging issues, containerization errors, and resource exhaustion—and provides actionable code fixes or configuration changes to get pipelines back on track. This skill is essential for researchers and engineers who need to minimize downtime in high-performance computing (HPC) and distributed environments.
Key Features
01Troubleshooting for containerized execution environments including Docker and Singularity
02Resource allocation optimization for OOM and job timeout failures
03Dependency graph validation to identify circular or missing references
04Pattern-based diagnosis for file staging and replica catalog issues
050 GitHub stars
06Automated log analysis using pegasus-analyzer and shell utilities
Use Cases
01Resolving container image pulling or command-path failures in remote jobs
02Debugging 'No such file or directory' errors in distributed file systems
03Fixing argument mismatches between workflow generators and execution wrappers
