Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

How Do You Find Traces of Log Deletion?

Deleted logs leave their own evidence: clearing events, gaps in record sequence numbers, truncated binary files, and recoverable records in unallocated space and memory.

Illustration for the cybersecurity term Log Deletion Traces
Home / Blog / Cybersecurity Terms / How Do You Find Traces of Log Deletion?
17 November 2026

How Do You Find Traces of Log Deletion?

How Do You Find Traces of Log Deletion?

An intruder who clears the logs usually believes they have removed the evidence. In practice they have created a new category of it. Deletion is an event, and on most systems the act of deleting leaves something behind — an entry recording the clearance, a discontinuity in numbering, a file whose size no longer matches its history, or a copy somewhere the attacker could not reach.

Windows: Start With the Obvious Entries

Clearing the Security log writes event ID 1102 to that log. Clearing the System or Application log writes event ID 104. Both survive the clearance itself, because they are the first record written to the newly emptied file, and both record the account that performed the action. Finding one is close to conclusive.

Finding none does not mean the logs are intact. More careful tooling removes individual records rather than whole files, a capability demonstrated publicly by the eventlogedit module in the DanderSpritz framework leaked in 2017. That approach leaves no clearance event, but it does leave gaps: every EVTX record carries a sequential EventRecordID, and the numbers should increase without interruption within a file. A jump from 48,112 to 48,140 is a finding, and a scripted check across every collected log file costs nothing.

Other signals worth sweeping for:

Linux: Size, Sequence, and Sealing

The binary login records — wtmp, btmp, and utmp — are commonly truncated to zero rather than edited, so compare their size and modification time against the system’s uptime and the activity you can see elsewhere. Selective editing is possible and leaves records that no longer align on expected boundaries, which utmpdump will expose.

Text logs under /var/log can be edited line by line, so look for timing inconsistencies instead: an authentication log with no entries during a window in which the web server log shows activity, rotated files whose numbering skips, or a file whose inode change time is later than its modification time, which indicates something altered it in place. Where systemd journal Forward Secure Sealing is enabled, verification will report tampering outright — it does not prevent deletion, but it removes the deniability.

Recovering What Was Removed

Deleted log data is frequently still present. EVTX records have a recognisable structure and can be carved from unallocated space, the page file, and volume shadow copies, with tools such as EVTXtract built for exactly that. Memory often contains recent events that the service had not yet flushed. Volume shadow copies and backups routinely hold a pre-deletion version of the whole file — the standard check that most teams forget under pressure.

These techniques sit alongside the broader anti-forensic techniques an intruder may use, and they are the reason deletion is rarely as effective as it appears.

The Control That Makes All of This Unnecessary

Forward logs off the host as they are written. A collector or SIEM holding a copy makes local clearance an indicator rather than an obstacle, and comparing the central record against the local file tells you exactly which entries were removed and when. For cloud audit trails, write to storage with object lock so that even an administrator cannot delete history.

Then alert on silence. A host that stops sending events is either switched off or being worked on, and both deserve attention. At Privia Security, we check for log integrity gaps at the start of every engagement in our investigation and incident response services, because an absence discovered on day one changes how the rest of the investigation is scoped.

FAQ – Frequently Asked Questions

Which Event ID Shows a Windows Log Was Cleared?

Event ID 1102 in the Security log and event ID 104 in the System or Application log. Both are written as the first entry in the newly emptied file and include the account that performed the clearance, so they survive the action they record. Their absence does not prove the logs are complete, because records can be removed individually instead.

How Do I Detect Selective Record Deletion?

Check the sequence. Every record in an EVTX file carries an incrementing identifier, and gaps within a single file indicate that entries were removed rather than aged out. Script the check across all collected logs, compare against a forwarded copy on a collector where one exists, and treat any discontinuity around the incident window as a priority finding.

Can Deleted Log Entries Be Recovered?

Often. Windows event records have a recognisable structure and can be carved from unallocated space, the page file, and volume shadow copies with tools built for the purpose. Memory may hold recent events not yet written to disk, and backups or shadow copies frequently contain the whole pre-deletion file, which is the quickest route when it exists.

What Stops Attackers Destroying Logs in the First Place?

Getting the data off the host immediately. Forwarding events to a collector or SIEM as they are generated means local clearance destroys only one copy, and the difference between the two sets shows precisely what was removed. Pair it with immutable storage for cloud audit trails and alerting on hosts that stop reporting.

Related Terms

You May Be Interested In These