Destroying evidence after an intrusion.
What gives it awayA clear with no matching change record, particularly on a server or domain controller, and particularly outside working hours.
ATT&CK T1070.001Nothing matches that yet. Tell us what you were looking for and it goes on the list.
Windows Security Log · Event 1102
Someone wiped the Windows Security log. There is almost no legitimate reason to do this on a production machine, and it is one of the few events that is a finding entirely on its own. Whatever was in the log is gone.
Also written as 1102Event ID 1102The audit log was clearedSecurity log cleared
The record of security activity on this PC was deleted. People occasionally do this to tidy up Event Viewer, but if you did not, it is a serious sign — erasing the log is one of the first things done to hide an intrusion.
This is why forwarding logs off the host matters: 1102 tells you evidence was destroyed but not what it said, unless you already shipped it elsewhere. The event survives the clear because it is written immediately afterwards. Treat it as act-now and pull the forwarded copy before anything else.
Sanitised. Addresses come from the RFC 5737 documentation ranges.
The audit log was cleared.
Subject:
Security ID: CORP\jbrooks-adm
Account Name: jbrooks-adm
Domain Name: CORP
Logon ID: 0x3E9A11 Everything else in the log line is context.
Most of the time it is one of these.
What gives it awayA clear with no matching change record, particularly on a server or domain controller, and particularly outside working hours.
ATT&CK T1070.001What gives it awayA clear shortly after account creation, group membership changes, or unusual privileged sign-ins on the same host.
ATT&CK T1070.001What gives it awayClears across several machines in the same window, which no legitimate process does.
ATT&CK T1070.001SecurityEvent | where EventID == 1102 | project TimeGenerated, Computer, ClearedBy = SubjectUserName, SubjectLogonId | order by TimeGenerated desc SecurityEvent | where EventID == 1102 | join kind=leftouter (SecurityEvent | where EventID == 4624 | project SubjectLogonId = TargetLogonId, IpAddress, LogonType) on SubjectLogonId | project TimeGenerated, Computer, SubjectUserName, IpAddress, LogonType Get-WinEvent -FilterHashtable @{LogName='Security'; Id=1102} | Select-Object TimeCreated, MachineName, Message index=wineventlog EventCode=1102 | table _time dest Account_Name Because Windows writes it as the first entry in the newly emptied log, immediately after the clear completes. It is the one record that deliberately survives the operation, which is exactly what makes it useful.
Occasionally — during a machine rebuild, when handing over a test system, or while troubleshooting a problem that needs a clean log to observe. What separates those from an attack is that they are planned, documented, and performed by someone who will confirm it when asked. An unexplained clear on a production server should be treated as an incident.
Not from the machine itself — the clear is destructive and there is no undo. You can recover it only if the events were already forwarded somewhere else, which is the strongest practical argument for centralised log collection. Forensic recovery from disk is occasionally possible but unreliable and not something to plan around.
Pull the forwarded copy of that host's logs from your collection platform before anything else, because that is now the only record of what happened. Then identify the account from the 1102 event, find where it signed in from, and treat the machine as untrusted until you can explain the clear.
Last reviewed 28 August 2026