I kept getting blamed for outages that weren't mine. So I built a tool to fight back

The villain origin story December 2024. 11 PM. I'm on the couch. buzzes. "Hey, the AI feature is broken." I check our dashboards. Everything's green. Our servers are fine. Our database is fine. Our CDN is fine. OpenAI is down. Not us. OpenAI. Their status page? Still showing "All Systems Operational." It took them over an hour to even acknowledge it. By then I'd already gotten 14 messages from users who thought we broke something. Two weeks later - same thing. OpenAI down again. 4+ hours. Same dance. Same blame. I started to notice a pattern.