Naive retries cut success from 55.4% to 41.5% under correlated failures, and 1 of 113 production configs randomizes delay.

Mehan and Saluja audited 200 open-source Python microservice projects. Explicit retry logic is detected in 11.5%, though their own false-negative audit puts true prevalence near 41%. Among detected projects 60.9% have at least one configuration with no backoff, and exactly one of 113 production configurations randomizes its delay (arXiv 2608.25403). In simulation at 100 trials per strategy, a naive retry policy under correlated failure does worse than not retrying at all. Jitter is a one-line change and almost nobody has made it.