Years ago, in a galaxy far, far away, my colleagues in support, Keith Elmore and Robert Dorr, built a tool called ostress.exe. For decades it transformed the way we reproduced customer issues for SQL Server. The engineering team used it, and up until now I’ve used it for demonstrations, books, and examples. Its magic was simple: ostress could simulate many concurrent users running SQL queries or scripts.
I decided to rethink ostress for today’s SQL world, and the result is sqlsimtools, now in public preview at https://aka.ms/sqlsimtools. It’s a free, open-source suite of workload simulation tools for SQL Server 2019, 2022, and 2025, Azure SQL Database, Azure SQL Managed Instance, and SQL database in Microsoft Fabric. I’ve been building it for some time and using it at events and in demonstrations. Now I want everyone to have it.
The public preview includes:
- sqlsim: a native, ODBC-based command-line tool for Windows (x64 and ARM64) and Ubuntu Linux.
- sqlpysim: a Python version of sqlsim built on Microsoft’s mssql-python driver, for Windows and Ubuntu Linux.
- sqlsim Studio: a self-contained local browser app that configures, runs, monitors, replays, and compares workloads using sqlsim or sqlpysim. It’s available today on Windows x64.
I’m also looking for Private Preview testers for macOS, and for sqlsim Studio on Linux and Windows ARM64. If you’d like to help, request Private Preview access.
I’ll be writing more posts on different parts of the tools, but here are my top 10 cool things you can do with sqlsimtools:
- Run any SQL query, script, or mixed workload
- Create your own benchmarks, load tests, and stress tests
- Configure, run, monitor, and compare workloads visually with sqlsim Studio
- Replay captured SQL Server activity
- Monitor workload performance live or after execution
- Run before-and-after experiments with detailed performance comparisons
- Test SQL Server, Azure SQL, and SQL database in Microsoft Fabric with flexible authentication
- Use native or Python tools on Windows, Linux, and macOS (Private Preview)
- Automate testing with JSON results and live event streams
- Build and run workloads with AI agents such as GitHub Copilot, Claude Code, Codex CLI, and Cursor
Get started today at https://aka.ms/sqlsimtools . The easiest way to start is the all-in-one sqlsimtools package, which includes sqlsim Studio and both command-line tools. You can also download just the tool for your platform, or clone the repo and build it yourself. On Windows, unblock the ZIP before you extract it, because the tools aren’t code signed.
sqlsimtools is an open-source project, not an official Microsoft product, and isn’t covered by Microsoft support. For questions and ideas, join the conversation in GitHub Discussions (microsoft/sqlsimtools · Discussions · GitHub), and report bugs in GitHub Issues (Issues · microsoft/sqlsimtools). During Public Preview I’m not taking Pull Requests yet but file issues and as the project progresses, we will incorporate the PR process.
Bob Ward
Microsoft