Google Cloud · 2026 Edition
A complete preparation guide written by Google Cloud-certified engineers. Covers the exam format,all 4 blueprint domains, a week-by-week study plan, and proven tips for passing first time.
3–5 months
Prep time
Advanced
Difficulty
60
Exam questions
720/1000
Pass mark
Exam code
PCDE
Full name
Google Professional Cloud Database Engineer
Vendor
Google Cloud
Duration
120 minutes
Questions
60 items
Passing score
720/1000 (scaled)
Domains covered
4 blueprint domains
Recommended experience
3+ years of database administration or development experience; familiarity with GCP data services
Typical prep time
3–5 months
The Professional Cloud Database Engineer certification validates the ability to design, create, manage, and troubleshoot Google Cloud databases. It is the credential for data engineers and architects building database solutions on GCP.
Job roles this opens
Domain percentage weights are not currently available for this exam. The checklist below is still useful for planning your study.
Weeks 1–3
Database Selection: Cloud SQL, Spanner, Bigtable, Firestore, Memorystore, AlloyDB
Tip: Database selection by use case is the most tested skill on PCDE. Know: Cloud SQL (managed MySQL/PostgreSQL/SQL Server, regional, max ~30TB), AlloyDB (PostgreSQL-compatible, 100x faster analytics than standard PostgreSQL, HA across zones), Cloud Spanner (global scale, strong consistency, unlimited horizontal scaling), Bigtable (NoSQL, single-digit millisecond reads at petabyte scale, time series/IoT/AdTech).
Weeks 4–6
Database Implementation: schema design, migrations, Cloud SQL Auth Proxy, connection pooling
Tip: Cloud Spanner schema design: Spanner uses interleaving to co-locate child rows with parent rows, which avoids cross-node joins. Know how parent-child table interleaving works and when it improves performance. Also know that Spanner auto-manages splits based on key ranges — a monotonically increasing primary key (like timestamp) creates hot spots.
Weeks 7–9
Database Migration: Database Migration Service, Datastream, AlloyDB Omni
Tip: Database Migration Service (DMS) supports continuous migration with minimal downtime. Know the source-to-destination combinations: MySQL/PostgreSQL to Cloud SQL, PostgreSQL to AlloyDB, Oracle to PostgreSQL (with conversion). Datastream is used for real-time change data capture (CDC) from databases to BigQuery, GCS, or Pub/Sub.
Weeks 10–13
Database Operations: monitoring, performance tuning, backups, HA/DR, compliance
Tip: Cloud Spanner Query Insights and Cloud SQL Query Insights both provide query-level performance data: P50/P95/P99 latency, execution plan visualisation, and lock contention metrics. Know which tool to use for performance investigation in each database service.
AlloyDB for PostgreSQL is a relatively new GCP service and is tested more heavily than its market share would suggest — Google is promoting it as a PostgreSQL-compatible alternative to Aurora. Know that AlloyDB has a disaggregated storage layer, a columnar engine for analytics (enabled per query with SHOW GENERATED COLUMNS), and built-in caching.
Bigtable row key design is critical for performance. Know that Bigtable stores rows in sorted order by row key and that queries by row key prefix are fast while scans across the entire table are slow. A good Bigtable row key combines multiple fields to support the most common access patterns (e.g. device_id#timestamp for IoT data).
Firestore in Native mode provides offline support for mobile/web clients and real-time listeners. Datastore mode is for server-side applications that need the Datastore API (e.g. migrating from Cloud Datastore). Know that you cannot change modes after a Firestore database is created.
Cloud SQL high availability uses a primary instance and a standby instance in a different zone in the same region. Failover is automatic — RPO is near-zero (synchronous replication to standby), RTO is typically 60 seconds. Know the difference between HA (same region, different zone) and Read Replicas (can be cross-region, asynchronous).
Memorystore for Redis cluster mode (Cluster tier) provides horizontal sharding across multiple Redis nodes for datasets that exceed a single-node memory limit. Know the difference between Memorystore for Redis (standard and cluster tiers) and Memorystore for Redis Cluster (separate product, OSS Redis Cluster protocol).
Apply everything in this guide with adaptive practice questions, detailed answer explanations, and domain analytics.
Deep-dive explanations of the key topics tested on PCDE — with exam key points and common misconceptions.