What's changed: Created Professional Cloud Database Engineer Chapter 4 (Domain 4 "Deploy HA": provisioning and testing = HA provisioning/IaC automation/failover drills/Cloud Monitoring; multi-regional replication and read replicas = regional resilience/read-write separation/continuous replication-lag monitoring/Spanner native multi-region).
4.2Multi-regional replication and read replicas
Understand setting up multi-regional database replication, deploying and scaling read replicas, separating reads from writes, and continuously monitoring the overall highly available setup.
To balance availability and scale, use cross-region replication and read replicas appropriately, designing the write and read paths.
4.2.1Multi-regional replication and read replicas
For regional resilience, replicate data to another region via multi-regional replication (Spanner is natively multi-region). To handle read load, deploy/scale read replicas and design read/write separation (writes to the primary, reads to replicas). Read replicas can aid availability too, but differ in purpose from HA auto-failover. Map "scale reads = read replicas" and "regional resilience = multi-regional replication."
4.2.2Continuously monitoring the setup
Continuously monitor multi-region/read-replica setups: watch each replica’s replication lag, read-replica load balancing, and failover readiness with Cloud Monitoring. High lag raises the risk of data loss on failover and reading stale values from read replicas. Map "continuously monitor replica lag and load."
Common: requirement → means. E.g., "distribute read load" = read replicas (read/write separation); "replicate for regional resilience" = multi-regional replication; "limit data loss on failover" = monitor/reduce replication lag; "native global strong-consistency multi-region" = Spanner.
Watch the mix-ups: (1) Read replicas (read scaling) and the HA standby (auto-failover) differ in purpose. (2) Read replicas may return stale values due to lag (use primary if strong consistency is required). (3) Multi-region adds latency and cost—choose by requirements.
4.2.3Section summary
- Regional resilience = multi-regional replication (native in Spanner)
- Scale reads = read replicas + read/write separation (distinct from HA standby)
- Continuously monitor replication lag and load with Cloud Monitoring
Sign in to track progress — Log in.
Quick check
(just a quick review)Q1. To distribute read load while concentrating writes on the primary, which best fits?
Q2. To replicate data to another region for regional resilience, which is best?
Q3. Which correctly contrasts read replicas and the HA standby?
Q4. What should you be careful about when reading from read replicas?
Q5. Which metric should you continuously monitor in multi-region/read-replica setups?
Q6. Which RDB natively offers a global, strongly consistent multi-region configuration?
Keep track of your progress
The full study guide is free to read. Sign up free to practice with the question bank, track what you have read, review your mistakes, and highlight passages.