GCP Database Engineer study plan

Compare Cloud SQL, Spanner, AlloyDB, Bigtable, Firestore, and BigQuery on consistency model, scaling axis, and HA topology, and record the trade-offs in a decision matrix

Design a reference architecture for a read-heavy web tier that separates read replicas from the primary and states the failover behavior explicitly

Configure a Cloud SQL high-availability instance and trigger a manual failover to observe the outage window

Explain Spanner's TrueTime and Paxos quorum model well enough to justify when multi-region strong consistency is worth the cost

Provision a Spanner multi-region instance and verify external consistency by running a read-write transaction that spans regions

Size a Cloud SQL instance and set connection limits, then demonstrate the effect of a connection pooler on max_connections exhaustion

Design a sharding strategy for Bigtable that avoids hotspotting on sequential row keys and document the key-design rationale

Compare regional vs multi-region availability SLAs and map each to a stated recovery objective

Deploy a Cloud SQL instance with private IP inside a custom VPC and verify connectivity from a Compute Engine client without public exposure

Configure IAM database authentication and a least-privilege service account for an application connecting to Cloud SQL

Set up automated backups, point-in-time recovery, and a retention policy, then restore to a specific timestamp

Write a runbook that describes how to promote a replica and repoint the application connection string

Design a polyglot persistence solution that routes transactional writes to Spanner and analytical reads to BigQuery without dual-write inconsistency

Compare change data capture options across Cloud SQL, Spanner, and Firestore and select one per source with justification

Define a data consistency contract that states which stores are system of record and which are derived

Evaluate Datastream or a Pub/Sub-based pipeline for streaming changes from an operational store into BigQuery

Build a migration assessment that classifies schema objects, stored procedures, and vendor-specific SQL features by conversion difficulty

Choose between Database Migration Service, dump-and-restore, and logical replication for three distinct source scenarios and justify each choice

Plan a minimal-downtime cutover with a replication lag check, a freeze window, and a rollback trigger

Identify and remediate schema incompatibilities such as unsupported data types, case sensitivity, and sequence behavior

Execute an end-to-end migration of a sample schema and dataset into Cloud SQL using Database Migration Service and validate row counts and checksums

Set up continuous replication and measure replication lag under a simulated write load

Perform a dry-run cutover, then execute the cutover and verify application queries against the new target

Document post-migration validation queries that prove data integrity beyond row counts

Tune query performance on Cloud SQL using Query Insights and add an index that measurably reduces a slow query's latency

Configure read pools and route read traffic away from the primary, then measure the throughput change

Design a disaster recovery topology that meets a stated RPO and RTO and identify the single points of failure it removes

Set budget alerts and choose committed use or edition sizing that matches the designed capacity

Synthesize the earlier designs into one end-to-end solution that spans operational and analytical stores with a defined migration path

Audit every design against the four domains and close the weakest gap with targeted review

Complete a timed set of scenario-based practice questions and analyze every miss by domain and root cause

Write concise justifications for the most common trade-off decisions so they can be recalled under time pressure