To help database teams answer that question before users start filing support tickets, Microsoft has introduced sqlsimtools (also referred to in reporting by Petri as SQLSimTools), a free, open-source suite of SQL workload simulation and analysis tools. Announced in a Microsoft Tech Community post on October 14, 2026, by Microsoft principal architect Bob Ward, the toolkit is now available in public preview.
While the release gives Windows and Linux database professionals a modern way to generate, replay, and compare workloads, it also comes with important preview boundaries and support caveats that enterprise IT teams need to understand before dropping it into their testing pipelines.
Rethinking a Classic: From ostress.exe to Modern SQL Environments
If you have spent any meaningful amount of time troubleshooting Microsoft database performance over the last two decades, you have probably crossed paths with ostress.exe, a venerable command-line utility included in the RML Utilities suite. As Ward noted in the Microsoft Tech Community announcement, former Microsoft support colleagues Keith Elmore and Robert Dorr built ostress.exe decades ago to simulate concurrent user connections running SQL queries or scripts, and it quickly became a staple for reproducing customer bugs, running engineering tests, and driving technical demonstrations.
However, database estates look very different today than they did when on-premises servers ruled the data center alone. Teams now straddle hybrid environments, cloud platform-as-a-service (PaaS) offerings, automated CI/CD pipelines, and Python-heavy data engineering workflows. Ward explained that he set out to rethink ostress.exe for modern SQL architectures, resulting in the sqlsimtools public preview.
According to Microsoft's announcement, the new open-source suite explicitly supports workload simulation across:
- SQL Server 2019, SQL Server 2022, and SQL Server 2025
- Azure SQL Database
- Azure SQL Managed Instance
- SQL database in Microsoft Fabric
That explicit inclusion of SQL Server 2025 and SQL database in Microsoft Fabric—alongside established Azure SQL tiers—makes the toolkit particularly relevant for organizations evaluating upgrades to Microsoft's newest data platform releases.
Inside the Toolkit: Three Components for Different Workflows
Rather than forcing every administrator and developer into a single interface, sqlsimtools splits its capabilities across three core components, as detailed by Petri and Microsoft's announcement:
| Component | Type | Underlying Tech & Role | Supported Platforms (Public Preview) |
|---|---|---|---|
| sqlsim | Native CLI | An ODBC-based command-line workload generator for running queries, scripts, and mixed workloads. | Windows x64, Windows ARM64, Ubuntu Linux |
| sqlpysim | Python CLI | A Python-based simulation tool built on Microsoft's mssql-python driver for Python-centric environments. | Windows, Ubuntu Linux |
| sqlsim Studio | Local Browser App | A self-contained local web interface for building, running, monitoring, replaying, and comparing workload tests. | Windows x64 |
It is worth noting a couple of practical distinctions here. First, while Petri describes sqlsim Studio broadly as a web-based interface, Microsoft's technical documentation specifies that it runs as a self-contained local browser application that orchestrates sqlsim or sqlpysim under the hood, rather than a hosted cloud service. Second, platform availability is not yet uniform across all three tools: while the command-line utilities already span Windows and Ubuntu Linux (with native Windows ARM64 support on sqlsim), sqlsim Studio is currently limited to Windows x64 in public preview. Microsoft is actively seeking private-preview testers for macOS support, as well as for bringing sqlsim Studio to Linux and Windows ARM64.
Core Capabilities: Benchmarking, Replay, and AI Agent Integration
What can database teams actually do with sqlsimtools once it is unpacked? Based on Microsoft's announcement and Petri's coverage, the toolkit focuses on repeatable, empirical verification before changes hit production:
- Targeted Load and Stress Testing: Teams can execute a single SQL query, a broader SQL script, or a complex mixed workload to create repeatable benchmarks, load tests, and high-concurrency stress tests.
- Workload Replay: Instead of guessing what production traffic looks like, administrators can replay captured SQL Server activity against a target test environment to observe how it handles real-world query patterns.
- Before-and-After Experiments: Using
sqlsim Studio, engineers can monitor performance live during execution or review it afterward, running structured before-and-after comparisons to see whether an index change, configuration tweak, or version upgrade improved throughput or introduced a regression. - Automation and Modern Auth: The tools support flexible authentication methods across on-premises and cloud targets and output structured JSON results alongside live event streams, making it straightforward to wire performance gates into automated DevOps pipelines.
- AI-Assisted Workflows: Rather than baking a proprietary AI service into the simulator itself, Microsoft designed
sqlsimtoolsso developers and DBAs can build and execute workloads using external AI coding agents—specifically naming GitHub Copilot, Claude Code, Codex CLI, and Cursor.
The Broader Context: Why Cloud Benchmarking Remains Tricky
sqlsimtools enters a space where community-driven tools have long done heavy lifting. For example, Microsoft Learn's Data Exposed series previously highlighted WorkloadTools, an independent open-source project created by Data Platform MVP Gianluca Sartori for capturing, analyzing, and replaying SQL workloads across on-premises and cloud environments.
As Microsoft Learn's coverage of WorkloadTools points out, lifting and shifting an application to the cloud looks easy on paper, but benchmarking PaaS environments like Azure SQL is notoriously tricky because cloud abstractions change what telemetry and system internals are exposed compared to a bare-metal Windows Server. Having a fresh, open-source toolkit authored inside Microsoft's SQL architecture group adds another welcome option for DBAs trying to bridge the gap between on-premises SQL Server and Azure or Fabric targets.
Still, as Petri rightly cautions, no simulation tool is a magic crystal ball. Based on general database administration principles, a benchmark is only as trustworthy as the workload you feed it and the environment you run it against. If your test database has a fraction of the data volume, different storage latency, or an unrepresentative mix of read and write queries, even the slickest replay engine won't prevent a production surprise.
Practical Adoption Caveats for Windows Administrators
Before you grab the download for your Windows workstation or jump box, Microsoft's announcement includes several important operational caveats that every sysadmin and security team should factor into their plans:
- Not an Official Microsoft-Supported Product: Even though
sqlsimtoolswas created by a Microsoft principal architect and announced on the official SQL Server blog, Microsoft explicitly states that the suite is not an official Microsoft product and is not covered by standard Microsoft support contracts. Support and feedback happen through GitHub Discussions and GitHub Issues, and the project is not yet accepting community pull requests during its public preview phase. - Unsigned Binaries on Windows: Microsoft distributes an all-in-one ZIP package containing
sqlsim Studioand both CLI tools (alongside options to download individual tools or clone and build from source on GitHub). However, Microsoft warns that the prebuilt binaries are not code-signed, meaning Windows users must manually unblock the downloaded ZIP archive in File Explorer properties before extracting it to avoid Windows security blocks. In strict enterprise environments enforced by AppLocker or Windows Defender Application Control (WDAC), administrators will need to inspect the repository and handle execution policies in a sandboxed test environment accordingly.
Summary and Key Takeaways
Microsoft's sqlsimtools public preview brings a modernized, open-source successor to the classic ostress.exe workflow, giving SQL Server, Azure SQL, and Microsoft Fabric teams a practical toolkit (sqlsim, sqlpysim, and sqlsim Studio) for stress testing, workload replay, and before-and-after performance comparisons. While the lack of formal Microsoft support and unsigned Windows preview binaries mean you should treat it as a community-grade testing utility rather than an enterprise-supported platform component, it offers Windows and Linux database professionals a versatile new way to validate migrations and upgrades before pushing changes live.
How are you currently stress-testing your SQL Server 2022, SQL Server 2025, or Azure SQL migrations? Drop into the WindowsForum.com database and enterprise IT discussions to share whether you'll be taking sqlsimtools for a spin in your test lab.