SnowPro Advanced Architect study plan

Diagram the three Snowflake architecture layers (storage, compute, cloud services) and explain which layer owns metadata, query parsing, and transaction locking

Trace the lifecycle of a query from client to storage, naming each component that touches the plan

Compare virtual warehouse types (standard, Snowpark-optimized, adaptive) and select the right one for three stated workloads

Explain micro-partitioning, clustering depth, and how pruning decisions are made against a filter predicate

Describe how Snowflake handles concurrency, resource contention, and warehouse queuing under multi-cluster scale-out

Configure a multi-cluster warehouse with stated min/max clusters and justify the scaling policy chosen

Set up a role hierarchy that models a three-tier organization and validate inheritance with test queries

Compare authentication methods (key-pair, SSO/SAML, OAuth) and pick one per use case

Write network policies and explain how they interact with private connectivity

Apply column-level, row-level (secure views, row access policies), and masking policies to a sample table and verify enforcement per role

Audit query history and access history to identify unused grants and anomalous access patterns

Read a query profile and identify the dominant cost node, spillage, and pruning gaps

Choose between clustering, search optimization, and materialized views for three given access patterns

Diagnose a slow query by comparing bytes scanned against partitions pruned

Tune a warehouse for a stated latency target by adjusting size, auto-suspend, and statement timeouts

Explain caching layers (result cache, local disk, metadata) and predict which queries will hit which cache

Design a data protection scheme using Time Travel, Fail-safe, and object tagging together

Classify data by sensitivity and map each class to a masking or tokenization policy

Configure data retention periods and explain the cost implications of longer retention

Set up a secure data share with a reader account and confirm that consumer-side access works without data copying

Explain how Snowflake Marketplace listings, listings with auto-fulfillment, and data clean rooms differ in governance model

Build a task graph with a root task, dependent tasks, and a stream-driven trigger

Choose between streams, dynamic tables, and Snowpipe for three ingestion scenarios

Describe exactly-once and idempotency guarantees when reloading from a stage

Configure Snowpipe with notification-based auto-ingest and validate file-load ordering and error handling

Set up a stage, file format, and COPY command with validation mode and error thresholds

Compare Snowpipe, Snowpipe Streaming, and batch COPY on latency, cost, and exactly-once behavior

Handle schema evolution during ingestion using schema detection and explain the tradeoffs

Design a recovery path for partially failed loads using COPY history and load metadata

Attribute credit consumption to warehouses, storage, and cloud services using account usage views

Set up resource monitors at account and warehouse level with stated thresholds and actions

Right-size warehouses and quantify savings from auto-suspend and auto-resume settings

Revisit architecture-layer questions from week 1 and close every gap in your notes

Identify the top three cost drivers in a sample account and propose a remediation per driver

Complete a timed practice set across all eight domains and score each domain separately

Re-teach your two weakest domains from scratch using documentation only

Re-run the week 1 architecture diagram and correct every component you cannot explain

Assemble a one-page cheat sheet of the 20 facts you keep forgetting