The vendor’s own documentation for each objective of the exam, free to use.
Security principles and recommendations for the AI and ML perspective in the Well-Architected Framework.
System design part of Architecture Framework has been updated to version 2.0, enabling better security, compliance, reliability, operations, and cost- and performance-optimization.
Explains configuring customer-managed encryption keys for BigQuery, covering Cloud KMS setup, key location matching, service account requirements, and CLI and SQL options.
Well-Architected Framework security recommendations for implementing security by design.
Overviews the security, privacy, and compliance pillar of the Google Cloud Well-Architected Framework, covering core principles, shared responsibility, and product focus areas.
Explores AI-native database architecture, showing how transactional, analytical, and vector data support agent memory, retrieval, and knowledge graph reasoning.
Guides estimating and controlling BigQuery costs, comparing on-demand and capacity pricing models and the dry-run and quota controls available.
Describes BigQuery custom query quotas for limiting on-demand query usage per project and per user, including how quotas reset, how to change them, and the errors raised.
Explains Bigtable app profiles, describing the standard and Data Boost types, their routing and priority options, and the default profile created with each instance.
Explains Bigtable replication, covering consistency guarantees, cluster placement, conflict resolution, and how app profile routing policies affect replicated reads and writes.
Describes how Cloud SQL high availability works, including primary and standby instances, synchronous replication, failover and failback behavior, and cost and SLA considerations.
Covers Cloud SQL MySQL replication, describing the replica types, their configuration constraints, and how replicas follow the primary instance.
Describes the Cloud Storage agent skills available for installation, what each skill covers, and the commands used to add them to a development environment.
Describes testing recovery from data loss, covering success criteria, backup verification, restore testing in non-production, and backup scheduling against recovery objectives.
Describes graceful degradation as a reliability design approach, covering load shedding, throttling, and comparisons with circuit breaking under high traffic.
Explains Google Cloud's unified analytics data platform architecture, covering BigQuery storage, compute, federation, and lakehouse components for large-scale data work.
Describes Google Cloud deployment architectures across single-zone, multi-zone, and multi-region setups, with availability tradeoffs to help choose a reliability design.
Provides guidance to help you assess the storage requirements of your cloud workload, understand the available storage options in Google Cloud, and design a storage strategy that provides optimal technical and business value.
Explains the horizontal scalability principle in Google Cloud's reliability pillar, comparing horizontal and vertical scaling and the role of stateless design and managed instance groups.
Explains observability for detecting failures, covering metrics, logs, and traces, and the monitoring and alerting practices used to spot under- or over-provisioned resources.
Well-Architected Framework reliability principle about testing for recovery from failures.
Provides principles and recommendations to help you design, deploy, and manage reliable workloads in Google Cloud.
Explains setting reliability targets in Google Cloud, covering error budgets, prioritizing critical components, and weighing recovery objectives against cost.
Explains the resource redundancy principle for high availability, covering replication across zones and regions, automatic failover, and status monitoring.
Explains how to perform single reads in Spanner outside transactions, comparing strong and stale reads and the related read lease and access control considerations.
Explains Spanner timestamp bounds for reads, comparing strong, bounded staleness, and exact staleness modes and the role of version retention.
Explains defining reliability from user-experience goals, covering which user-facing metrics to measure and how to use tracing to find latency issues.
The convergence of data lake and data warehouse means that users can now access the data regardless of the infrastructure limitations, in where or how data is stored.
There are numerous design patterns that can be implemented when processing data in the cloud; here is an overview of data pipeline architectures you can use today.
Describes how to design Dataflow pipelines for flexibility and efficiency by choosing worker machine types, streaming features, and cost-related pipeline options.
Describes multicloud database deployment patterns, disaster recovery models, and database categories to help architects choose database options spanning multiple clouds.
Compares ELT and ETL data integration approaches for BigQuery, naming transformation tools and export options for moving data out of BigQuery.
Surveys methods for loading data into BigQuery, including batch loads, streaming, transfers, federated queries, and replication, to help choose the right ingestion approach.
When embarking on a large, multi-year data center migration to Google Cloud, itâs helpful to think of the project in phases.
Data and database migration is key for your business’s cloud and growth strategies. Learn how Google Cloud can help you scale.
Discusses how to set up and execute the database migration process, including failure scenarios.
Gives an overview of Database Migration Service, comparing continuous and one-time migrations, homogeneous and heterogeneous types, and its security features.
Moving your existing data is a key component of cloud migrations. Learn more about modernization and migration approaches for data workloads.
Describes how to design, implement, and validate a plan to migrate from Amazon EC2 to Compute Engine.
Get a guide for transferring your data to Google Cloud and learn what tools are available.
Describes multicloud database deployment patterns, disaster recovery models, and database categories to help architects choose database options spanning multiple clouds.
Explains the continuous and one-time migration flows in Database Migration Service for MySQL, including the dump and change-replication phases.
Shows how to promote a Database Migration Service migration job with gcloud, checking the asynchronous operation and prerequisites before cutover.
Provides steps to create Pub/Sub Cloud Storage subscriptions, covering required IAM roles, service account permissions, bucket prerequisites, and output file naming.
Introduces Cloud Storage subscriptions for Pub/Sub, explaining how they write messages to buckets, batching and retry behavior, and when Dataflow is preferable.
There are numerous design patterns that can be implemented when processing data in the cloud; here is an overview of data pipeline architectures you can use today.
Lists best practices for writing Dataflow pipeline code, covering DoFn design, fault tolerance, dead-letter handling, and side inputs versus CoGroupByKey.
Explains how to plan Dataflow pipelines around service level objectives, isolation of data and resources, external service calls, and Pub/Sub usage.
Explains Pub/Sub exactly-once delivery, covering supported subscription types, acknowledgment ID behavior, regional requirements, and when duplicates can occur.
Describes BigQuery change data capture ingestion, covering the required pseudocolumns, primary key rules, API requirements, and table limits for applying upserts and deletes.
Explains how to configure a BigQuery destination for Datastream, covering dataset organization and the merge and append-only write modes.
Describes the metadata columns and write behavior Datastream applies when streaming change data into BigQuery tables, for merge and append-only modes.
Explains primary and foreign key constraints in BigQuery, how to create, add, and drop them, and which metadata views expose them, noting they are not enforced.
Describes the BigQuery Storage Write API over gRPC, covering required permissions and scopes, and the stream types and write semantics for ingesting data.
Learn how to define, build, and compile your machine learning pipelines in Gemini Enterprise Agent Platform Pipelines.
Explains how Dataflow detects bottlenecks in pipelines, the categories and causes reported, and troubleshooting steps for BigQuery sink and Kafka source issues.
Explains Dataflow streaming modes and how to enable them with service options or gcloud flags, including changing modes on existing jobs.
Automate, monitor, and govern your machine learning (ML) systems in a serverless manner
Overviews the serverless deployment of Managed Service for Apache Spark, covering batch workloads, interactive sessions, supported Spark languages, and tooling integrations.
Build, deploy, and troubleshoot Airflow pipelines in minutes using Google Cloud Data Agent Kit and natural language.
Explains Pub/Sub message ordering, covering ordering keys, regional publishing rules, throughput limits, and how redelivery and dead-letter topics affect order.
This article reviews three approaches to building a streaming data pipeline on Google Cloud, using Pub/Sub and BigQuery.
Shows how to diagnose slow or stuck Dataflow streaming jobs using freshness, backlog, and latency metrics, parallelism charts, and worker logs.
Lists best practices for writing Airflow DAGs in Cloud Composer, covering scheduling, retries, naming, and metadata settings to keep environments healthy.
Basic overview of deploying and running pipelines in Cloud Data Fusion.
Learn how to configure and deploy orchestration pipelines.
Lists Google Cloud services with brief descriptions and terms-of-service notes, useful for identifying which offerings suit compute, storage, and management needs.
Describes the managed Apache Airflow service, its Python-based DAG pipelines, integrations, and configuration and pricing considerations for workflow orchestration.
Introduces best practices for implementing machine learning (ML) on Google Cloud, with a focus on custom-trained models based on your data and code.
Explains MLOps practices for continuous integration, delivery, and training of ML systems, and compares manual level 0 with automated level 1 pipeline maturity.
Outlines preparing production ML pipelines, covering hardware and quota planning, and what each pipeline should monitor for skew, drift, and outages.
Explains Pub/Sub seek and snapshots for replaying or purging messages, covering snapshot lifetime, seeking by time versus snapshot, and configuration requirements.
Shows how to build PyTorch ML workflows as Vertex AI Pipelines using KFP components, pipeline decorators, artifacts, and prebuilt Google Cloud pipeline components.
Walks through running an Apache Airflow DAG in Managed Airflow Gen 2 using gcloud, listing required IAM roles and service account setup steps.
Explains how to schedule and trigger Airflow DAGs, defining the date concepts, start dates, catchup behavior, and retry settings that govern DAG runs.
Teaches how to write Airflow DAGs in Managed Airflow, covering task definition, dependency syntax, and the operators and tools available on workers.
Lists Bigtable schema design guidance and size and count limits for row keys, rows, cells, column families, and tables to help plan table layouts.
Cost-optimization recommendations and best practices
Explains Cloud Storage Bucket Lock, describing how bucket retention policies work, when they can be locked, and how they interact with object versioning and metadata.
Explains Cloud Storage soft delete behavior, including retention settings, how to restore deleted objects and buckets, and how restored objects are handled.
Data lakes and data warehouses differences include data storage, users, tasks, and accessibility. Learn more about how these systems work.
Explains how to manage partitioned BigQuery tables, including inspecting partitions through metadata views and setting or removing partition expiration with SQL, bq, and API options.
Introduces Spanner schema and data model concepts such as tables, primary keys, and interleaving, with guidance on parent-child deletion behavior.
Guides avoiding hotspots in Spanner schema design through key choice, UUID and hashing functions, and index strategies for high write-rate tables.
Describes clustered tables in BigQuery, covering eligible column types, clustering column rules, and how clustering affects block pruning and bytes processed.
Check out the migration framework and architecture we suggest for moving your data warehouse, like Teradata, to Google Cloud BigQuery.
Overviews the BigQuery Migration Service, listing supported source systems and the SQL translation and migration tools available for moving workloads.
Explains how to use nested and repeated fields in BigQuery schemas, with guidance on STRUCT and ARRAY types, UNNEST, and performance tradeoffs versus flattened tables.
Explains how to write queries that enable partition pruning in BigQuery, covering pseudocolumns, filter expression patterns, and the require partition filter option.
Explains BigQuery partitioned table types, including time-unit and integer-range partitioning, and the rules for how values and boundaries are assigned to partitions.
Explains BigQuery time travel and fail-safe data recovery, covering how to configure the recovery window and how deleted tables and row-level policies are affected.
Explains how to create, manage, and drop vector indexes in BigQuery, including index type choices, required permissions, and column and distance requirements.
Overviews BigQuery vector search capabilities, covering embedding generation functions, search and similarity functions, vector indexes, and how these operations are billed.
Cloud data warehouses bring flexibility and scale, but affect people as well as technology. Hereâs how DBAs can adapt and use modern data warehouses.
A data warehouse is an enterprise system used for the analysis and reporting of structured and semi-structured data from multiple sources.
Learn decision points necessary in creating a modern, unified analytics data platform built on Google Cloud.
Explains how to manage partitioned BigQuery tables, including inspecting partitions through metadata views and setting or removing partition expiration with SQL, bq, and API options.
A high-level architecture to build a borderless open data lakehouse that establishes a highly governed, secure pipeline from raw borderless silos to AI and agentic driven actions.
The convergence of data lake and data warehouse means that users can now access the data regardless of the infrastructure limitations, in where or how data is stored.
Make sure your data lake is working for your business and see how to use tools like Apache Kafka to migrate from on-prem.
Data lakes and data warehouses differences include data storage, users, tasks, and accessibility. Learn more about how these systems work.
Shows how to build an open data lakehouse with Delta Lake on Dataproc, querying the tables from Presto through Dataproc Metastore and manifest files.
Data-driven transformation starts with a modern data platform, built on Google Cloud
Describes the architecture and components of a data mesh that's deployed in Google Cloud.
Describes design considerations for building data products in a data mesh.
There are numerous design patterns that can be implemented when processing data in the cloud; here is an overview of data pipeline architectures you can use today.
Explains BigQuery authorized views, authorized materialized views, and their location, permission, and quota considerations for restricting access to source data.
Introduces BI Engine for in-memory acceleration of BI queries, covering supported integrations, reservations, and which tables and query patterns benefit from acceleration.
Explains BigQuery column-level access control with policy tags and data governance tags, including required roles and permissions for reading and modifying protected columns.
Introduces column-level data masking in BigQuery, covering masking rules, policy tag and data governance tag options, and the data types each rule supports.
Introduces BigQuery materialized views, covering their types, refresh behavior, freshness settings, and restrictions on definitions and data operations.
Describes how to query materialized views and what conditions determine whether BigQuery applies incremental updates or falls back to the base query.
Introduces row-level security in BigQuery, describing how filter expressions and grantee lists work, who can create policies, and how policies interact with other access controls.
Shows how to create, update, and drop row-level access policies in BigQuery, with guidance on required permissions and how multiple policies interact.
Describes design considerations for building data products in a data mesh.
Explains how to share Data Studio reusable and embedded data sources, comparing access levels and what each permission allows.
Looker's options for sharing data.
Explains how to create, manage, and drop vector indexes in BigQuery, including index type choices, required permissions, and column and distance requirements.
Overviews BigQuery vector search capabilities, covering embedding generation functions, search and similarity functions, vector indexes, and how these operations are billed.
Describes how to prepare and curate data for machine learning, covering sufficient sample sizes, missing values, leakage, skew, and granularity concerns.
Describes how to prepare tabular training data for classification and regression models, covering size limits, target column rules, split methods, and supported data sources.
Easily and securely exchange valuable datasets and analytics assets in BigQuery across any organizational boundary with a fully managed service.
Describes BigQuery Analytics Hub data sharing, covering exchange and listing types, publisher and subscriber roles, and how linked datasets work.
Explains how to share Data Studio reusable and embedded data sources, comparing access levels and what each permission allows.
Exchange and monetize historical and real-time market data at scale securely and easily.
Looker's options for sharing data.
Describes BigQuery custom query quotas for limiting on-demand query usage per project and per user, including how quotas reset, how to change them, and the errors raised.
Cost-optimization recommendations and best practices
Provides guidance to optimize the resource usage of your cloud workloads for sustainability.
Shows how to create, apply, and manage autoscaling policies for Managed Service for Apache Spark clusters, including configuration fields and CLI commands.
Explains secondary worker types for Managed Service for Apache Spark clusters, covering their behavior, default settings, and CLI flags for configuring them.
Offers guidance on preparing Google Cloud resources for peak traffic events, covering right-sizing, autoscaling, pre-warming, monitoring, and alerting practices.
Use Application Design Center to streamline application design and create reusable templates. Learn how teams can manage a catalog of approved templates to accelerate development.
Explains BigQuery partitioned table types, including time-unit and integer-range partitioning, and the rules for how values and boundaries are assigned to partitions.
Describes Google Cloud's data cloud agents for engineering, science, database, and analytics tasks, with conversational analytics and related use cases.
Describes the architecture of an agentic data science system on Google Cloud, covering root and specialized agents, data sources, deployment, and security recommendations.
Introduces database migration concepts and principles, covering migration types, replication, active-passive and active-active modes, and consistency requirements.
Describes an agent-based approach to automating data governance in BigQuery using column-level lineage, covering descriptions, glossary terms, policy tags, and its CLI and dashboard.
Overviews Managed Service for Apache Airflow, comparing Gen 1, Gen 2, and Gen 3 versions and describing DAGs, environments, and their core components.
Describes troubleshooting common DAG problems in Managed Airflow, covering logging, parsing and timeout settings, sync delays, and related configuration options.
Introduces Google Cloud Workflows, a serverless orchestration service, covering workflow definitions, executions, deployment, and regional behavior for service coordination.
Explains how to run BigQuery queries, comparing job types and destination options, and lists the IAM roles and permissions required to run them.
Explains BigQuery slots as the compute units behind query execution, contrasting on-demand and capacity-based pricing and how slot allocation and reservations work.
Describes the Cloud Billing onboarding checklist and how organizations, folders, projects, labels, and billing accounts relate in the Google Cloud resource hierarchy.
Offers guidance on planning GKE clusters for large workloads, covering node scale, Compute Engine quotas, and API request limits to raise.
Explains how Google Cloud resources are organized into organizations, folders, and projects, and how identifiers and isolation between projects work.
Discusses how to plan your hybrid and multi-cloud strategy.
Learn how application performance monitoring helps you monitor, troubleshoot, and optimize your applications for peak performance and reliability.
Guides estimating and controlling BigQuery costs, comparing on-demand and capacity pricing models and the dry-run and quota controls available.
Offers best practices for optimizing BigQuery query compute performance, covering column selection, partition filtering, and query structure choices.
Explains how BigQuery slot reservations and assignments organize compute by job type and resource hierarchy, with notes on edition limits and billing behavior.
Explains how Streaming Engine works in Dataflow, including SDK support, enablement options, and constraints for updating or restarting streaming pipelines.
Explains using Gemini Cloud Assist investigations to diagnose failed or slow Managed Service for Apache Spark clusters and jobs, covering setup and the report layout.
Google Cloud Observability suite is designed to monitor, troubleshoot, and improve cloud infrastructure and application performance.
Well-Architected Framework operational excellence recommendations to manage incidents and problems with cloud workloads.
Explains the VM observability dashboards in the Google Cloud console, covering basic versus Ops Agent metrics, fleet views, and customizing dashboards.
Describes testing recovery from data loss, covering success criteria, backup verification, restore testing in non-production, and backup scheduling against recovery objectives.
Explains how to update a running Dataflow job, covering update options, transform name mappings, validate-only graphs, and the REST method for runtime parameter changes.
Explores disaster recovery scenarios for data on Google Cloud, comparing storage, disk, and BigQuery recovery options and their tradeoffs for backup and restore planning.
Describes Google Cloud's process for identifying, reporting, and resolving data incidents, including response stages, severity factors, notification channels, and AI involvement.
Explains secondary worker types for Managed Service for Apache Spark clusters, covering their behavior, default settings, and CLI flags for configuring them.
Explains observability for detecting failures, covering metrics, logs, and traces, and the monitoring and alerting practices used to spot under- or over-provisioned resources.
Well-Architected Framework reliability principle about testing for recovery from failures.
Argues for monitoring and alerting on user-facing symptoms rather than underlying causes, discussing alert design, actionability, and the continued role of cause monitoring.