Challenges of Managing Stateful Applications in Kubernetes
summary
Claims that running databases on Kubernetes is now production-ready and used by large deployments like CockroachDB, Crunchy Data, Mongo Atlas, backed by concrete tooling: CSI drivers to resize volumes, custom resource definitions and operators to automate live CPU/memory scaling, failover, backups and Prometheus-based monitoring. Tradeoffs versus VM or cloud-managed databases include lower latency and local SSD performance, higher operator burden, vendor lock-in risks and a cited cost shock (speaker reports up to $667,000 for a single managed DB). Appscode’s qdb protocol and a simple UI are presented as working ways to tame that complexity.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:07Stateful apps challenges
- 4:30Database management approaches
- 9:23VM-based database drawbacks
- 13:16Cloud provider trade-offs
- 16:48Vendor limitations
- 17:31Databases in Kubernetes benefits
- 20:00Self-managed complexity
- 21:32CRDs vs operators
- 23:41KubeDB introduction
- 24:43UI tools for deployment
- 25:29Q&A
related talks
35mStateful Apps On Kubernetes - Bring Them On!
45mDatabases on Kubernetes: Why you should care
08mLightning Talk: Should You Run Your Database on Kubernetes
47mFrom a Database in Container to DBaaS on Kubernetes
45mFrom Laptop to Cloud: Developing Cloud Native Applications with Containerized Databases
59mA DevOps State of Mind: Managing Microservices + Databases with Kubernetes
25mPostgres on Kubernetes for the Reluctant DBA
45mBest Practices for Running Databases on Kubernetes