Podcast
Tech stories for people who run things. Solo narration. Deep research.
A misconfigured public bucket, a third-party CDN that was not caching, and twenty-seven days of egress. The bill that arrived was $84,000. What happened next.
A 2am page for a slow database. The dashboards say everything is fine. He restarts the connection pool. Then he fails over. Three hours later, he finds out what was actually wrong.
At 11:39 AM Eastern, October 4th 2021, a Facebook engineer ran a routine command. Within minutes, every Facebook service was gone. Not down. Gone. The story of BGP, DNS, and the safety mechanism that turned a connectivity failure into total erasure.