How to identify what is slowing down a Windows PC
Use measured workloads and built-in resource tools to locate a bottleneck.
- Updated
- September 18, 2026
- Reading time
- 8 min read
Use measured workloads and built-in resource tools to locate a bottleneck. Use the procedure as a controlled diagnostic process and preserve a safe way back.
Create a baseline
Choose one repeatable slow task, noting whether the delay occurs before sign-in, when apps start, or during ordinary use. Save work before a normal restart if needed, then let startup and known maintenance settle rather than using a fixed waiting period.
Compare a local task with an online one. A single slow website with responsive local apps needs a network or service comparison, while several slow local tasks justify checking system resources.
Read Task Manager
Press Ctrl+Shift+Esc. Select More details if shown, then inspect Processes while reproducing the delay. Sort CPU, Memory, Disk, and Network individually and note the process that stays busy at the same time; a brief spike is not enough to assign blame.
CPU: distinguish an expected update, scan, or calculation from an app that remains busy after its task ends. Memory: compare process use with Performance > Memory, especially available and committed memory. Cached memory is reusable; a high percentage alone does not establish a leak.
Disk: use Performance to identify the affected drive and compare active time and response time. High active time need not mean high transfer speed. If disk activity rises while available memory is low, paging may contribute; if one app reads many files, its workload may instead explain the delay.
Use Resource Monitor
Press Windows + R, enter resmon, and press Enter. Its CPU, Memory, Disk, and Network tabs provide more detail. Select the checkbox beside a suspected process to narrow the associated activity, and clear the selection before comparing other processes.
On Disk, examine file paths and read/write activity to distinguish document access, indexing, or other work. On Memory, hard faults mean pages had to be fetched from disk, not that RAM is physically defective. Persistent faults with low available memory and a matching slowdown support investigating memory pressure.
On Network, compare the named process and transfer activity with the slow online task. These observations show correlation, not proof: do not disable services or terminate unfamiliar processes just because they appear near the top.
Confirm the cause
Save work and close one known nonessential app normally, or let an identified legitimate background task finish. Repeat the same task with the same files and compare both responsiveness and the suspected resource. Restore changes that did not help.
If the delay persists without the suspected pressure, investigate another boundary, such as the app, network, or storage device. Stop for overheating, drive errors, or crashes; protect files and provide process names, timing, and relevant observations to support instead of continuing stress tests.