Diagnostic ETL Logs
Windows Diagnostic ETL Logs are Event Trace Log files written under ProgramData by the Connected User Experiences and Telemetry pipeline and related diagnostic tracing sessions. They store rotating .etl trace data such as provider events, performance timings, upload state, and reliability diagnostics so Windows can analyze issues and submit telemetry, but old sessions can accumulate after failed uploads, tracing errors, or long runtimes. Kudu removes stale ETL log files from this diagnostic trace store without touching personal files, accounts, saved settings, or application data.
Why clean Diagnostic ETL Logs?
- Accumulated .etl trace sessions in ProgramData can grow into hundreds of megabytes or more, and the user notices unexplained system disk usage in storage reports even though no personal files were added
- Orphaned diagnostic traces left behind after telemetry upload failures keep old session data on disk, so the folder keeps refilling and free space slowly shrinks between reboots
- Corrupted or partially written ETL files can make diagnostic tools and support scripts take longer to parse logs, which shows up as slow troubleshooting scans or errors opening trace data
- Very large retained trace logs increase antivirus and indexing work on every scan, and the symptom is extra background disk activity from system processes with no obvious foreground app causing it
- After Windows feature updates, older diagnostic trace files may no longer be useful for current troubleshooting, yet they still occupy space and make it harder to identify fresh logs from a new problem
- Continuous tracing over long periods can leave many rolled ETL segments behind, and users typically notice the Diagnostic folder growing again after crashes, update failures, or repeated reliability issues
Cache paths Kudu targets
Windows
%ProgramData%/Microsoft/Diagnosis/ETLLogs |
Common questions about Diagnostic ETL Logs
Download Kudu and reclaim your disk space.
Available on Windows, macOS, and Linux. No account required, no feature gates, no telemetry without consent. All cleaning targets are open source and community-auditable.