Google Cloud Cloud Database Engineer

Under-sized databases fail and over-sized ones waste money. You estimate compute, memory, storage and connection needs from workload data, plan growth, and choose between scaling up, out and with replicas.

• Estimating CPU, memory and storage
• IOPS, throughput and connection limits
• Growth forecasting
• Vertical scaling, read replicas and sharding
• Right-sizing from monitoring data
• Capacity planning for peaks

You build A sizing plan for one database with growth forecast and scaling triggers.

Many database outages begin as connection problems. You manage connection limits with pooling, choose secure connection paths, and fit connections to serverless and container platforms.

• Connection limits and what exhausts them
• Session pooling and transaction pooling
• Private IP and the Cloud SQL Auth Proxy
• Connections from Cloud Run, GKE and functions
• Retry and failover handling in clients
• Connection security and certificates

You build A connection design with pooling and tested behavior under load.

Availability costs money, so design it to the need. You set recovery time and recovery point targets, choose zonal, regional and multi-regional layouts, and plan replication across Regions.

• Recovery time and recovery point targets
• Zonal and regional high availability
• Read replicas and cross-region replicas
• Multi-region configurations in Spanner
• Failure domains and what they take down
• Costing availability options

You build A high availability and disaster recovery design with targets and cost for one database.

A failover plan that is never tested is a guess. You provision highly available instances with code, set up cross-region replication, and rehearse failover and recovery so the real behavior is known.

• Provisioning with gcloud and Terraform
• Configuring high availability and replicas
• Setting up cross-region replication
• Forced failover and promotion of replicas
• Disaster recovery drills
• Recording measured recovery times

You build A provisioned high availability database with a drill report of measured recovery times.

Databases hold what attackers want most. You control access with IAM and database roles, protect connections and data with encryption, and audit who did what.

• IAM roles for database administration
• IAM database authentication
• Database users, roles and privileges
• Customer-managed encryption keys
• Network isolation for databases
• Audit logging and review

You build An access and encryption design for a database, with an audit log review.

Slow queries and locks cause most database incidents. You monitor key metrics, find expensive queries, read execution plans, and fix indexing and contention problems.

• Metrics and alerts for databases
• Query Insights and slow query analysis
• Reading execution plans
• Indexing strategy
• Locks, deadlocks and long transactions
• Connection and resource exhaustion

You build A performance investigation of a slow workload with the fix and its measured effect.

Backups define how much you can lose. You design backup schedules and retention to meet recovery targets, use point-in-time recovery, copy backups across Regions, and test restores.

• Automated and on-demand backups
• Retention and storage cost
• Point-in-time recovery
• Exports, imports and logical backups
• Cross-region backup copies
• Restore testing and verification

You build A backup and recovery plan with a documented point-in-time restore test.

Databases need steady upkeep. You tune size and settings to cut cost without hurting performance, schedule maintenance windows, and automate recurring tasks and upgrades.

• Right-sizing and committed use discounts
• Configuration and parameter tuning
• Maintenance windows and version upgrades
• Automating routine tasks with schedulers and functions
• Storage growth and cleanup
• Reviewing cost by database

You build A maintenance and cost plan with an automated task running on a schedule.

Migrations succeed or fail in the plan. You assess source databases, choose a strategy and tool for each, set downtime limits, and prepare for compatibility differences.

• Assessing source databases and dependencies
• Homogeneous and heterogeneous migration paths
• Downtime tolerance and cutover options
• Choosing a migration tool
• Compatibility and schema conversion planning
• Fallback planning

You build A migration plan for three source databases with strategy, tool and downtime for each.

A live migration moves a running system. You set up continuous replication with Database Migration Service, convert schema and code for different engines, validate data, cut over, and keep reverse replication ready for fallback.

• Continuous replication with Database Migration Service
• Schema and code conversion across engines
• Validating data during and after migration
• Cutover steps and timing
• Reverse replication for fallback
• Migrating databases hosted on other clouds or on premises

You build A rehearsed migration with continuous replication, validation checks and a fallback path.

Design and run one database estate. You select engines, size and secure them, build availability and backup, tune them, and migrate one database with a plan you can defend.

• Gathering workload requirements
• Selecting and sizing databases
• Designing availability, security and recovery
• Migrating one database
• Presenting the estate and its runbooks

You build A documented database estate with selection, availability, recovery and migration plans.