Resource Governor: Reducing Dastardly Behaviour in SQL Server 2025
Proposed session for SQLBits 2026TL; DR
SQL Server 2025 allows bigger, busier workloads, and not all of them behave. Learn how Resource Governor can control noisy neighbours, when it genuinely helps, and when it quietly makes things worse.
Session Details
SQL Server 2025 increases the memory limits available to Standard Edition, making it far more practical to consolidate multiple databases and workloads onto a single instance. For many organisations, this means stacking more systems with different usage patterns, priorities, and performance characteristics onto shared infrastructure.
That consolidation makes performance problems harder to manage, monitor, and diagnose. Issues are less likely to come from a single obviously bad query and more likely to emerge from the interaction between otherwise reasonable workloads competing for memory, CPU, and TempDB. Symptoms often appear at the instance level, while the root causes remain difficult to isolate.
Resource Governor, now available in Standard Edition, is frequently suggested as a way to control these situations. It has a mixed reputation because it is often used to contain problems rather than understand them. While it can be effective at limiting resource consumption between workloads, it can also mask underlying design issues or make troubleshooting more complex if applied without a clear goal.
In this session, we will examine what Resource Governor actually controls, how it affects memory grants, CPU scheduling, and TempDB usage, and how it behaves in heavily consolidated environments. We will focus on recognising when Resource Governor can genuinely improve stability, when it makes diagnosis harder, and when it is being used as a substitute for fixing the workload itself.
This session is not a configuration walkthrough or a feature tour. It is about using Resource Governor deliberately, with a clear understanding of what problem you are trying to solve, rather than hoping it will make a difficult system easier to live with.
That consolidation makes performance problems harder to manage, monitor, and diagnose. Issues are less likely to come from a single obviously bad query and more likely to emerge from the interaction between otherwise reasonable workloads competing for memory, CPU, and TempDB. Symptoms often appear at the instance level, while the root causes remain difficult to isolate.
Resource Governor, now available in Standard Edition, is frequently suggested as a way to control these situations. It has a mixed reputation because it is often used to contain problems rather than understand them. While it can be effective at limiting resource consumption between workloads, it can also mask underlying design issues or make troubleshooting more complex if applied without a clear goal.
In this session, we will examine what Resource Governor actually controls, how it affects memory grants, CPU scheduling, and TempDB usage, and how it behaves in heavily consolidated environments. We will focus on recognising when Resource Governor can genuinely improve stability, when it makes diagnosis harder, and when it is being used as a substitute for fixing the workload itself.
This session is not a configuration walkthrough or a feature tour. It is about using Resource Governor deliberately, with a clear understanding of what problem you are trying to solve, rather than hoping it will make a difficult system easier to live with.
3 things you'll get out of this session
After attending this session, you will be able to:
Recognise when performance problems in consolidated SQL Server instances are caused by workload interaction rather than individual poorly written queries.
Decide when Resource Governor is an appropriate tool for isolating workloads, and when its use is likely to make diagnosis and troubleshooting harder.
Evaluate the impact Resource Governor has on memory grants, CPU scheduling, and TempDB usage in busy, multi-database environments.
Speakers
Richard Douglas's other proposed sessions for 2026
Couchbase for the SQL DBA - 2026
TempDB: The Bottleneck You Can’t See in the Execution Plan - 2026
Richard Douglas's previous sessions
Improving SQL Server Performance with Indexing
Indexing presents daunting challenges for even the most seasoned professionals, it offers countless options to choose from. With a little help you’ll see how to simplify indexing in your environment & improve
the overall performance of your apps
Investigate TempDB like Sherlock Holmes
Learn how to investigate the mysterious case of SQL Server performance problems in TempDB.
Alerting; Beyond the Basics
We'll look at events that you should be monitoring, how you can use various services & IoT enabled devices to alert you in ways that were never possible. You'll leave with a new outlook of how teams can be notified to fix problems faster than ever.
The Big Backup Theory
In this session, you'll gain invaluable guidance for optimizing your backup strategies. Using trace flags the session will cover how SQL Server manages its own backup options and how you can tune them to meet your SLAs.
Wait watchers-Gain Performance Increases Fast!
In this session you'll learn all about SQL Server wait statistics. These are statistics the database engine stores about the resources it is waiting on. Armed with this information you as a SQL Server professional can make better informed decisions.
The Day After Tomorrow; Why You Need to Baseline
Ensuring top performance is never easy. To maintain performance you need to make sure you’re monitoring the right things. How do you know if the figures you’re seeing are good or bad? Baseline comparisons can help.