Some of my background services occasionally fail due to transient issues, yet they never try again on their own. I’d like clean, dependable retry logic added so that any job that exits unexpectedly—or returns an error code I define—will be queued up and executed again automatically, with sensible back-off and a clear upper limit to avoid endless loops. Here’s what I need from you: • Review the current job runner and identify where retries belong. • Add configurable retry parameters (maximum attempts, delay strategy, failure thresholds). • Log each attempt with enough context to trace root causes later. • Provide a concise README explaining how to tune or disable the retry behaviour. The codebase is small and straightforward; you’ll have full access to a private repository as soon as we agree on the approach. Please mention any prior experience implementing exponential back-off, circuit breakers, or queue-based retries when you apply—seeing real-world examples will help me pick the right partner quickly. Looking forward to making these jobs bullet-proof together.