MySQL in the Cloud: DBaaS vs Kubernetes
summary
Compares Google CloudSQL auth proxy vs Percona Operator for MySQL on GKE with concrete tradeoffs: Cloud SQL gives managed failover (heartbeat detection ~60s, docs say <3 minutes but real cases can be longer) and easy auth via sidecar, while the operator provisions XtraDB Cluster with HAProxy or ProxySQL, supports anti‑affinity, autoscaling limits, and async replication channels for DR. Shows a cost example: six services with 3‑node clusters — CloudSQL ~ $250k vs running on Kubernetes roughly half that, and operators automate many DBA tasks.
AI-written summary · about
chapters (4)
related talks
36mCreate your Own MySQL-as-a-Service that Runs Anywhere Using Kubernetes Operators
23mRunning MySQL on K8s. Deployment, Scaling, Backups and advanced configuration
19mPrivate DBaaS: Why you Need it, How to Build It
41mDemocratizing MySQL: From Cloud Managed to Kubernetes Managed
27mDIY DBaaS with Percona Kubernetes Operators
50mThe hell of SQL DBs on Kubernetes versus the promise of operators
51mMySQL in the Cloud
44mFrom a database in container to DBaaS on Kubernetes