#BuildPropulsionLab at #DevoxxUK: @jbaruch with @ladymeyy
summary
Rollout expands the IP range for pods on Kubernetes and modifies the company networking stack at quter, deployed dev to prod by region; minutes after triggering the change production load balancers stop responding and DNS routes traffic to the wrong service. Recovery uses infrastructure-as-code to identify changes and route traffic to VMs as a quick rollback, restoring service. Concrete takeaways: always have a verifiable rollback plan, test the cluster and networking itself not only services, and maintain cross-team collaboration and communication.
AI-written summary · about
related talks
50mDevOps Project Failures - Lessons Learned and Strategies for Recovery
42mMastering Kubernetes: Challenges, Strategies, and Best Practices
52mDevops @Scale
16mKeynote: What a RUSH! Let’s Deploy Straight to Production!
23mDressing to the 9s
30mBreaking Barriers: Bringing Application Developers Closer To the CNCF
28mDevOpsDays 2026 Ignites Day 2
34mKubernetes from Dev to Prod at GoEuro