About this lab
PulseMetrics is a B2B SaaS analytics platform. Every account on the platform generates usage events (dashboard views, API calls, exports), and the data team needs that raw activity turned into something the business can trust and query cheaply.
In this lab you set up separate virtual warehouses for loading versus analyst workloads and cap spend with a resource monitor. You model accounts and subscriptions as permanent tables, and stage raw usage events in a transient table. You parse nested JSON usage events with dot notation and FLATTEN them into an analytics-ready fact table, then expose that data through a standard view, a materialized view for fast rollups, and a secure view that hides billing details from anyone but finance.
Finally, an engineer's bad DELETE wipes real usage data. You use Time Travel to recover it, UNDROP a table that was dropped outright, and then clone the entire schema into a dev environment so the next risky change can be tested without touching production.
Lab details
- Level: Beginner
- Cloud: Snowflake
- Domain: Data Engineering
- Duration: 100 minutes
What you will learn
- Create and size virtual warehouses for different workload types, and cap spend with a resource monitor
- Model data with permanent vs transient tables and understand the storage and Fail-safe tradeoff
- Query raw JSON event data with dot notation, bracket notation, and FLATTEN
- Build standard, materialized, and secure views for different consumers of the same data
- Recover from an accidental DELETE with Time Travel and bring back a dropped table with UNDROP
- Clone a schema instantly for a safe dev environment with zero-copy cloning
Concepts covered
- Virtual warehouses and warehouse types
- Resource monitors and cost governance
- Permanent vs transient tables
- Standard, materialized, and secure views
- Semi-structured JSON with VARIANT and FLATTEN
- Time Travel (BEFORE/AT), UNDROP, and zero-copy cloning
Technologies used
- Snowflake Virtual Warehouses
- Snowflake Resource Monitors
- Snowflake Table Types
- Snowflake Views
- Semi-Structured Data (VARIANT)
- Time Travel & Zero-Copy Cloning
Prerequisites
- Basic SQL (SELECT, JOIN, GROUP BY)
- A Snowflake trial account (Enterprise edition features like materialized and secure views are used)
Topics
- Snowflake
- Data Warehousing
- SQL
- Data Engineering
- Virtual Warehouses
- Views
- Time Travel
- Resource Monitors
