Advanced
Replication and HA Patterns
Operate a small HA footprint with clarity: replication slots, monitoring signals, and failover drills that respect maintenance windows. Labs favor transparency over magic failover buttons.
Duration: 4 weeks · 30 hours · Format: Cohort + tabletop day
Price (informational): ￥176,000
Request information
What is included
- Lag SLOs tied to business metrics
- Slot management and disk risk watchlist
- Controlled switchover runbook draft
- Read replica routing pitfalls
- Observability signals beyond replication lag
Outcomes
- Author a failover checklist your on-call can follow
- Define a lag alert that avoids noisy pages
- Identify two scenarios where reads should avoid replicas
Lead contact
Sora Nguyen
SRE lead for regional SaaS; builds tabletop exercises for data outages.
FAQ
We reference RDS and Cloud SQL patterns narratively, not as vendor labs.
Operators mentioned briefly; this is not a Kubernetes course.
No multi-region consensus algorithms; pointers only.
Participant notes
Tabletop day exposed our stale promotion lag metric. Follow-up template saved a week of arguing.
Client in manufacturing