Current Workflow
Follow Alice's PR from push to production
Payment retries rewritten — fixed 5s intervals replaced with exponential backoff. 3 files changed in src/payments/.
A colleague reviews the code for quality, but nobody checks which product features this touches, whether the docs need updating, or who should hear about it.
The change is live in production, but the PM doesn't know. Docs still describe fixed intervals, and three partner integrations depend on retry timing via webhooks.
A support ticket arrives: a partner's webhook handler assumed 5s intervals. Nothing linked Alice's PR to the partner webhooks, so nobody saw it coming.
Root cause: "We didn't know this change affected partner integrations." All the information existed — code, runtime, dependency graph — nobody connected them.
The bug wasn't expensive to fix. The expensive part was the two weeks nobody knew it was broken.