Stop Hand Tuning MySQL Teach Kubernetes Operator
summary
Percona’s MySQL Kubernetes operator adds an auto config Go library that computes config from pod CPU/memory limits and workload type and writes a ConfigMap, replacing the operator’s previous 3 tunings with 31 parameters (buffer pool, group replication, connections, redo/innodb settings). Measured result: a 4 CPU 16 GB pod does 454 TPS, doubling to 8 CPU 32 GB without autoconfig yields ~472 TPS, but with autoconfig reaches ~1,445 TPS — roughly 3x improvement. Enable it by flipping the feature, choosing a workload type, and setting the MySQL version.
AI-written summary · about
chapters (10)
- 0:00MySQL performance challenges
- 1:57Unoptimized MySQL scale
- 2:38Manual tuning flaws
- 3:43Auto-config in MySQL operator
- 5:56How auto-config works
- 8:30Key parameters adjusted
- 10:00Enabling auto-config
- 12:12MySQL version handling
- 13:20Auto-config logic
- 14:02Call to action
related talks
36mRight-Sizing and Auto-Scaling of MySQL Containers in Kubernetes
23mRunning MySQL on K8s. Deployment, Scaling, Backups and advanced configuration
17mWorkload-Based MySQL Tuning
16mWhether tuning MySQL configuration important?
26mHow MySQL Configuration Tuning Boosts App Performance
53mMySQL DevOps
43mOptimizing MySQL without SQL or touching my.cnf
36mRight-Sizing and Auto-Scaling of MySQL Containers in Kubernetes