Is Your Kubernetes Disaster Recovery Actually Ready?
summary
68% of organizations report data loss last year with an average cost of $4.5M and fewer than one third can answer Kubernetes DR specifics. Demos use Velero + Kopia to S3, Longhorn standby volumes, Argo CD/GitOps, kind and Kayak clusters to show a Git rebuild leaves PVCs empty until Velero restores, and the volume group snapshot (Kubernetes 1.36) eliminates inter-PVC snapshot races but remains crash not application consistent. Core gaps: no atomic multi-PV DB snapshot, no native cross-cluster failover, no standard application-unit DR, and almost nobody rehearses end-to-end restores.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 2:29Kubernetes recovery question
- 3:12RPO and RTO explained
- 4:35Kubernetes recovery layers
- 4:55Demo: Velero backup/restore
- 10:20Longhorn cross-cluster DR
- 12:49DR tool integration gaps
- 13:09DR architecture patterns
- 16:18Demo: GitOps + Velero rebuild
- 20:18Demo: Broken Kubernetes recovery
- 23:53Volume group snapshots
- 25:38Failover sequence gaps
- 26:42Measuring recovery time
- 27:03Production-ready DR setup
- 28:08From questions to actionable steps
- 29:10Key takeaways and closing remarks
related talks
35mDisaster Recovery for your Kubernetes Clusters
33mNavigating Disaster Recovery in Kubernetes and Crossplane: Strategies and Tools for Managing Stateful Workloads with GitOps for Platform Teams
15mWhen Clusters Collide: The Misadventures of Disaster Recovery in K8s
46mKubernetes Stateful Application Disaster Recovery with Zero Data Loss
01h 00mApplication resilience in Kubernetes... certainty even before your DRP?
06m⚡ Lightning Talk: Disaster Recovery - How IaCaC and Kubernetes Enables Cost Efficiency and Fast Recovery
31mKubernetes Backup and Migration Strategies with Velero
34mKubernetes Backup and Migration Strategies with Velero