Lucas and Luna dig into a 2024 postmortem from a mid-sized fintech that lost two hours of transaction processing because their internal API gateway — the one tool every microservice relied on — went down during a routine config change. They walk through what happened, why the 'just add another gateway' fix isn't as simple as it sounds, and how companies like Uber and Stripe have moved to sidecar proxies to avoid this exact failure mode. Specific numbers: the fintech normally handled 1,200 requests per second through that gateway, and the outage cost an estimated $340,000 in failed transactions and manual reprocessing. If you run any kind of service mesh or API gateway in production, this episode will make you rethink your failover strategy.

#APIGateway #SinglePointOfFailure #Microservices #Postmortem #Fintech #SidecarProxy #ServiceMesh #Stripe #Uber #EnvoyProxy #ResilienceEngineering #IncidentResponse #DevTools #API #Infrastructure #Business #BusinessPodcast #FexingoBusiness

Keep every episode free: buymeacoffee.com/fexingo