Overview
Added retry and repetition count tracking to an execution model, providing visibility into retry behavior and enabling better debugging of transient failures.
Problem Statement
Users had limited visibility into retry behavior:
- No way to see how many times a step retried
- Difficult to distinguish transient vs persistent failures
- No data for retry policy tuning
- Hard to debug intermittent issues
- Missing metrics for reliability analysis
Solution
Implemented retriesCount and repetitionsCount properties on execution steps, tracking both automatic retries and manual repetitions with full history.
Key Features
Automatic Retry Tracking
- Count of automatic retries performed
- Reason for each retry
- Duration between retries
- Final success/failure status
Historical Data
- Full retry history with timestamps
- Error messages for each attempt
- Success on which attempt
- Retry pattern analysis
Analytics Enabled
- Most common retry reasons
- Average retries per step type
- Success rate by attempt number
- Optimal retry delays
Results & Impact
- Visibility: Complete transparency into retry behavior
- Debugging: Faster incident resolution
- Tuning: Data-driven retry policy optimization
- Reliability: Identified patterns requiring fixes