This repository contains samples, demos, and benchmarking tools to help you get started with Snowflake Interactive Analytics scenarios — including interactive warehouses, interactive tables, and low-latency, high-concurrency workloads.
Looking for the Interactive Benchmark skill? Starting with Cortex Code (CoCo) CLI v1.2.18 and CoCo Desktop v1.22.1, the interactive-benchmark skill is available natively — no need to clone this repository. Just open CoCo CLI or CoCo Desktop and invoke the skill directly.
Full documentation available here: The interactive-benchmark skill
Follow the Getting Started guide for prerequisites and a step-by-step walkthrough.
Benchmark harnesses for Snowflake Interactive Warehouses, sharing one Python core (benchmarks/core/) for setup, execution, timing, validation and result storage. Each runs against standard, interactive and iceberg tables, on standard or interactive warehouses, at scale factors 1, 10, 100 and 1000. See benchmarks/README.md.
benchmarks/tpc-h/— copies TPC-H tables fromSNOWFLAKE_SAMPLE_DATAinto a local benchmark database and runs the 22 standard queries (original and modern rewrites). Seebenchmarks/tpc-h/README.md.benchmarks/ssb/— Star Schema Benchmark: derives the SSB star schema from the TPC-H sample data and runs the 13 SSB queries (original, modern and flat/denormalized variants). Seebenchmarks/ssb/README.md.
Complete sample applications that show interactive warehouses in real workloads.
example_apps/interactive-vs-standard/— "The Concurrency Test": a Snowflake-hosted Next.js app that runs the same retail queries on an XSMALL standard warehouse and an XSMALL interactive warehouse side by side, with 1 to 100 virtual users, and compares p95 latency, errors and queue time. Seeexample_apps/interactive-vs-standard/README.md.example_apps/postgres_sync/— the same analytical dashboard (Cube Core + Next.js, 32 tiles) running against Postgres and Snowflake, with a synthetic data generator and a connectivity check. Seeexample_apps/postgres_sync/README.md.
TPC-H interactive dashboard demo. A FastAPI + Chart.js single-page app that visualizes TPC-H data (KPIs, orders over time, segment/region breakdowns) with a Locust load generator for concurrent user simulation. Deployable to SPCS. See tpc-h-sample-dashboard/README.md for setup and usage.
A Cortex Code skill that benchmarks Snowflake Interactive Warehouse performance under concurrent load using any user-provided SQL query. The skill automates the entire workflow — from verifying that your tables and queries are suitable for interactive warehouses to deploying the benchmark infrastructure and generating reports. See interactive-benchmark/README.md for sample prompts, repository structure, and SPCS deployment details.
This repository includes three project-level CoCo skills that automate common workflows through conversational prompts in Cortex Code.
Benchmark any SQL query on interactive warehouses under concurrent load.
- Chains the
snowflake-interactiveskill to verify tables and queries are suitable for interactive warehouses - Deploys the benchmark API and Locust load test to Snowpark Container Services
- Reports whether the query meets its P95 latency goal under load
Usage: Ask something like "benchmark my query on an interactive warehouse".
Set up and run the TPC-H benchmark locally against interactive warehouses.
- Creates TPC-H tables at various scale factors (1, 10, 100, 1000)
- Runs the 22 standard queries and collects timing results (JSON + CSV)
- Supports both
originalandmodern(window functions, QUALIFY) query variants
Usage: Ask something like "set up the TPC-H benchmark at scale 10" or "run tpch queries against interactive".
Deploy the TPC-H dashboard demo and Locust load test to SPCS.
- Deploys FastAPI + Chart.js dashboard with real-time TPC-H visualizations
- Deploys a Locust load generator for simulating concurrent dashboard users
- Manages SPCS service lifecycle (deploy, status, update, teardown)
Usage: Ask something like "deploy the TPC-H dashboard to SPCS" or "check tpch dashboard status".
More samples and scenarios will be added over time.