Powerful Windows Performance Counter Monitoring

Continuous, centralized visibility into CPU, disk usage, memory, and application performance across all your Windows servers.

How do I monitor Windows performance counters across multiple servers?

Windows performance counters are built into the Microsoft operating system and expose a wide range of metrics. From CPU usage and disk read rates to application-specific data from SQL Server, IIS, and the .NET CLR. Every Windows machine already has this data; the question is how you collect it and turn it into action.

Paessler PRTG polls performance counter data continuously across local and remote Windows machines, stores it historically, and visualizes it in continuous (near real-time) graphs and dashboards. It reads both native operating system-level counters and counters registered by third-party applications, making it practical for mixed Windows operating system environments of any size.

PRTG monitors performance counters for: Windows Server 2008 R2 and later, IIS, SQL Server, .NET CLR, Windows Services, Azure-hosted Windows VMs, and any third-party application that registers counters with the Windows performance counter subsystem.

Download PRTG Trial

PRTG is compatible with all major vendors, products, and systems

What PRTG Does With Your Windows Performance Data

Always-On Visibility Across Every Windows Server

Perfmon works great for checking one machine when you already know where to look. Centralized monitoring scales that to your entire fleet, collecting data continuously whether someone is watching or not.

PRTG continuously polls your Windows servers and collects data centrally, then sends notifications when a counter value crosses defined thresholds. That way, you’re not dependent on someone watching dashboards.

Continuous polling across Windows devices on a configurable interval
All performance data collected and stored automatically
Centralized view of counter status across every monitored server simultaneously
Threshold-based alerts send notifications when values exceed defined limits
Remote monitoring using configured credentials (in some cases, specific sensors may require additional Windows services/settings)

Spot Trends and Plan Capacity With Historical Data

Server performance tells a story over time. Patterns like memory trending upward or disk latency increasing under load become visible when you have historical data. With that context, you understand what changed over time rather than only seeing the current state.

PRTG retains counter metrics historically and shows them in time-series graphs per sensor. That means you can review hours, days, or months of data to support troubleshooting, spot gradual bottlenecks, and make capacity decisions with observed usage, not guesswork.

Performance data retained historically (review hours, days, weeks, or months back per sensor)
Time-series graphs show trends in resource usage per counter and per device
Identify gradual degradation before it’s obvious to users
Support capacity planning and upgrade decisions based on actual usage data
Exportable reports for recurring review cycles or management visibility

See Application Performance Down to the Counter

CPU and memory tell you the operating system is under load, but not which application is driving the behavior. On Windows, SQL Server, IIS, and the .NET CLR expose their own performance counter objects.

PRTG can read performance counters registered on the Windows system, so SQL Server, IIS, and .NET CLR counters can be monitored alongside OS counters.

SQL Server: buffer cache hit ratio, batch requests/sec, page reads/sec
IIS: application pool health, worker process failures, request queue length
.NET CLR: exceptions per second, GC heap size, thread count
Third-party app counters registered to the Windows performance counter subsystem

See Why IT Professionals Trust PRTG

Start monitoring your infrastructure in minutes. No professional services, no complex configuration, no risk.

Free download

PRODUCT OVERVIEW

Stable, Lightweight Data Collection Out of the Box

By default, PRTG uses WMI to collect Windows performance data, which is the same underlying access method used across most Windows sensors. For environments running large numbers of Windows sensors, PRTG supports a hybrid mode that tries performance counters first and falls back to WMI only when needed. This reduces WMI overhead and helps keep probe performance stable as your sensor count grows.

WMI is PRTG's default collection method for Windows performance sensors
Optional hybrid mode: try performance counters first, fall back to WMI when unavailable
Hybrid mode is configured via Windows Compatibility Options on the parent device
When a sensor runs in fallback mode, PRTG retries via performance counters after 24 hours
Hybrid mode reduces WMI load and is recommended for environments with large numbers of Windows sensors

Start Monitoring What Matters Right Away

A standard Windows Server can expose hundreds of counter objects. PRTG's preconfigured sensors cover the ones that matter most, so your team can focus their experience where it counts.

PRTG’s Windows sensors cover critical OS counters, including disk usage, so you can start quickly. Auto-discovery can identify Windows devices and create relevant sensors automatically, and device templates reduce setup effort when adding new machines. When you need counters beyond the prebuilt set, PerfCounter Custom lets you add any registered counter.

Preconfigured sensors for core OS counters (CPU, memory, disk usage, network)
Auto-discovery for Windows devices to create sensors automatically
Device templates in auto-discovery to reduce manual setup
Extend monitoring to additional counters using PerfCounter Custom
Scales from a few servers to hundreds with manageable setup

How PRTG Monitors Windows Performance Counters

You have multiple different possibilities to monitor your environment with PRTG. This section covers the technical side of how Windows performance counter data is collected: collection methods, sensor types, and configuration concepts.

Agentless Counter Collection

PRTG reads performance counter data remotely from the probe, without requiring an additional agent on the monitored machine. The probe connects to the target Windows system using configured credentials.

Some sensors may require additional Windows services/settings on the monitored system, depending on the sensor and its access method.

Hybrid WMI Fallback

PRTG sensors use WMI by default. For sensors that support it, you can enable a hybrid mode in Windows Compatibility Options: PRTG will then try performance counters first and use WMI as a fallback if the counter is not available or readable. When running in fallback mode, PRTG retries via performance counters again after 24 hours. Sensors that are WMI-only continue using WMI regardless of this setting.

PerfCounter Sensor Types

PerfCounter Custom sensor: monitors a configured set of Microsoft operating system performance counters. You define counter objects/instances and add channels. It supports monitoring any counter registered in the Windows performance counter subsystem (native OS counters and those registered by third-party apps).

PerfCounter IIS Application Pool sensor: monitors an IIS application pool via Windows performance counters.

Standard Windows sensors such as Windows CPU Load use WMI by default to request monitoring data. If hybrid mode is enabled via Windows Compatibility Options, those sensors will try performance counters first and fall back to WMI when needed.

Alerting and Dashboards

Each monitored counter can be represented as channels. You configure upper and lower thresholds per channel; when values cross your thresholds, PRTG sends alerts using your chosen notification methods. Sensor data can be visualized in graphs and used in custom dashboards.

PowerShell Script Sensors

For scenarios not covered by native sensor types, PRTG supports PowerShell script sensors. A script can query performance counter data (for example using Get-Counter) and return values to PRTG for storage and graphing. This requires scripting knowledge and is typically for custom or advanced monitoring needs.

free downLoad

Gathering Windows Performance Data: Manual vs. PRTG

FEATURE	Without PRTG Without PRTG	With PRTG With PRTG
Multi-server monitoring	Without PRTG Manual, one machine at a time in perfmon	With PRTG Centralized — all Windows devices in one view, polled continuously
Historical data retention	Without PRTG Not retained by default; manual export required	With PRTG Stored automatically, viewable per sensor at any time
Threshold-based alerting	Without PRTG No alerting in perfmon	With PRTG Configurable per channel, with multiple notification methods
Application counter coverage	Without PRTG Requires manual counter selection per machine in perfmon	With PRTG Any registered counter accessible via PerfCounter Custom
Data collection method	Without PRTG WMI or manual perfmon access: neither scales across multiple machines or retains data automatically	With PRTG WMI by default; optional hybrid mode (performance counters first, WMI fallback) configurable via Windows Compatibility Options to reduce overhead at scale

free downLoad

“We strive to equip our systems with state-of-the-art technology to safeguard our educational practices for the future. Part of this includes ensuring that all our systems run smoothly at all times. On any given day, we rarely have time to keep an eye on all our systems. We therefore decided to monitor our school’s IT environment with a centralized network monitoring tool.”

Stefan Roschewitz, IT administrator

BBS Holzminden

With PRTG, we are now more proactive and have improved the quality of our services. Now, we have everything measured: power, air conditioning, tanks, surveillance cameras (CCTV), processors, memories, disks, switches, bandwidth, and network traffic. We can see what is happening and act before any problem arises.

Esbin Saúl Lázaro García, IT Infrastructure and Security Engineer

Hospital El Pilar

The reactivity, know-how, and technical solutions of Paessler support are outstanding in every situation. For me, no monitoring tool compares to PRTG.

Andreas Reimann, Senior Networking Communication Architect

Zurich Airport

Paessler PRTG Network Monitor licenses & pricing

Choose the PRTG Network Monitor subscription that's best for you.

License Name	Price	License Details	Get started	Pricing Details
PRTG 500	^$200	per month paid annually	Buy now Buy now	Enough to monitor multiple aspects of 50 devices
PRTG 1000	^$358	per month paid annually	Buy now Buy now	Enough to monitor multiple aspects of 100 devices
PRTG 2500	^$742	per month paid annually	Buy now Buy now	Enough to monitor multiple aspects of 250 devices
PRTG 5000	^$1,300	per month paid annually	Buy now Buy now	Enough to monitor multiple aspects of 500 devices
PRTG 10000	^$1,642	per month paid annually	Buy now Buy now	Enough to monitor multiple aspects of 1000 devices

Over 100,000 Customers Worldwide Love Paessler

Windows Performance Counter Monitoring: Frequently Asked Questions

What’s the difference between Windows performance counters and WMI?

Windows performance counters are a Windows performance data source, the same one perfmon uses. WMI is a broader management framework covering system inventory and configuration, plus performance data. For routine performance monitoring, performance counters are typically the more efficient and stable choice.

By default, PRTG uses WMI to collect Windows performance data. For environments with high sensor counts, PRTG supports an optional hybrid mode, configured via Windows Compatibility Options, that tries performance counters first and falls back to WMI when needed. This reduces WMI overhead without sacrificing coverage.

Which Windows performance counters should I monitor on a Windows Server?

For general health, typical starting points include:

Processor: % Processor Time
Memory: Available MBytes, Pages/sec
PhysicalDisk: Disk Read Bytes/sec, Avg. Disk Queue Length
Network Interface: Bytes Total/sec

Beyond that, role-based counters matter (SQL Server and IIS differ). PRTG’s preconfigured sensors cover core OS counters, and PerfCounter Custom lets you define application-specific counters.

Can I monitor performance counters on remote Windows machines?

Yes. PRTG collects counter data remotely using configured credentials, without requiring an agent on the monitored machine. Some sensors may require additional Windows access or services depending on the collection method.

Does PRTG support SQL Server and IIS performance counters?

Yes. SQL Server and IIS expose counters through the Windows performance counter subsystem. With PerfCounter Custom, you can monitor specific SQL Server counters. For IIS, the PerfCounter IIS Application Pool sensor provides predefined channel monitoring for application pool health and related counters.

Does this work for Azure-hosted Windows servers?

It works when the Azure VM is reachable from a PRTG probe over the network. PRTG still needs network access to query the Windows system, like any other monitored server.

How does PerfCounter Custom work?

The PerfCounter Custom sensor monitors a configured set of Windows performance counters. In PRTG, you add the sensor to a Windows device and configure the counter object, instance, and counter name. Each configured counter becomes a channel with its own thresholds and graphs. You can include multiple channels within one sensor.