From a Database in Container to DBaaS on Kubernetes
summary
Explains why databases move from Docker (good for dev/test but risky for prod) to Kubernetes using PersistentVolumes, StatefulSets and the Operator pattern; claims Kubernetes still needs careful setup for mission‑critical data but already runs “tens if not hundreds of thousands” of production DB nodes; cites PlanetScale sharding to containers under ~256 GB; highlights Percona’s offerings—Percona Distribution for MySQL/Postgres/Mongo, Percona Operators and PMM GUI—to deliver an open‑source DBaaS; warns RDS price premium recently approaches ~2x on newer instances.
AI-written summary · about
chapters (27)
- 0:00Introduction
- 1:40The drive to simplicity
- 4:05Database in Docker
- 5:54Database in production
- 7:53State of open source database
- 8:28Percona approach to Docker
- 9:16Day 2 operations
- 10:33Kubernetes and databases
- 15:20Where Kubernetes is helpful
- 16:46What Kubernetes can do
- 19:55Percona operators
- 20:48Unsolved problems
- 21:40DBaaS
- 22:13Current state of DBaaS
- 23:53Benefits of DBaaS
- 24:33Challenges of DBaaS
- 25:39Fully managed
- 26:50Cost premium
- 28:17Oracle history
- 29:53Cloud vs electricity
- 32:37Open source vs proprietary
- 35:13Our vision
- 35:39Work in progress
- 36:43DBaaS and open source
- 39:10Summary
- 40:05Simple words
- 40:28Questions
related talks
36mFrom a Database in Container to DBaaS on Kubernetes
44mFrom a database in container to DBaaS on Kubernetes
19mPrivate DBaaS: Why you Need it, How to Build It
25mAre Containers Ready for Production Databases?
34mDatabases on Kubernetes
45mDatabases on Kubernetes: Why you should care
26mChallenges of Managing Stateful Applications in Kubernetes
45mBest Practices for Running Databases on Kubernetes