On November 30th, GitHub experienced significant network issues leading to 18 minutes of complete unavailability, along with intermittent slow responses and errors throughout the day. The problems stemmed from a misconfiguration during a network upgrade aimed at improving scalability and performance, which inadvertently caused a bridge loop and revealed a bug in the aggregation switches. This bug prevented MAC addresses from being learned, leading to network saturation. Despite extensive troubleshooting and collaboration with the network vendor, the issue persisted until a workaround involving restarting core processes on the switches was implemented. GitHub plans to address the root cause with a software update and is committed to improving its infrastructure and incident response through enhanced monitoring, staging environments for testing, and more proactive support engagement. The company acknowledges the impact on users and expresses gratitude for their patience during the incident.