Auth Log Hunting
Hunt for credential abuse and lateral movement using authentication telemetry.
Last updated: February 2026Purpose and Scope
Authentication logs are among the most valuable data sources for detecting compromised credentials and lateral movement. This playbook covers hunting techniques for identifying unauthorized access, credential theft, and attacker movement across Windows, Linux, and cloud environments.
Prerequisites
- Log sources: Windows Security logs, Linux auth logs, Azure AD/Entra sign in logs, VPN logs
- SIEM access: Query capability across authentication data
- Baseline knowledge: Understanding of normal authentication patterns in your environment
- Asset inventory: Knowledge of critical systems and privileged accounts
Key Data Sources
Windows Authentication
- Event ID 4624: Successful logon
- Event ID 4625: Failed logon
- Event ID 4648: Explicit credential use (RunAs)
- Event ID 4672: Special privileges assigned (admin logon)
- Event ID 4768: Kerberos TGT request
- Event ID 4769: Kerberos service ticket request
- Event ID 4776: NTLM authentication
Linux Authentication
- /var/log/auth.log or /var/log/secure: SSH, sudo, su events
- Accepted/Failed password: SSH login attempts
- session opened/closed: User session management
- sudo commands: Privilege escalation events
Cloud Identity
- Azure AD sign in logs: Interactive and non interactive sign ins
- Azure AD audit logs: Account and group changes
- AWS CloudTrail: Console and API authentication
- Okta system logs: Authentication events and failures
Hunting Hypotheses
Credential Stuffing and Spraying
Attackers try many passwords against one account or one password against many accounts:
- Look for accounts with many failed logons in short period
- Look for single source IP failing against many accounts
- Identify spray patterns: same error, same time window, many targets
Splunk example:
index=wineventlog EventCode=4625
| stats count by src_ip, user, _time span=1h
| where count > 10
| sort -count
Impossible Travel
User authenticates from geographically distant locations in short time:
- Compare sequential logon locations for same user
- Flag authentications from different countries within hours
- Account for VPN and proxy usage in baseline
Unusual Logon Times
Legitimate users have patterns; attackers may not:
- Baseline normal working hours per user or group
- Flag after hours authentication to sensitive systems
- Watch for weekend access to systems typically idle
First Time Access
User or system accessing resource for first time may indicate compromise:
- First logon to a server by a user
- First remote access from a workstation
- New source IP for a user account
Lateral Movement Indicators
Pass the Hash / Pass the Ticket
Attackers use stolen credential material instead of passwords:
- NTLM authentication from unusual sources
- Kerberos ticket requests without corresponding TGT
- Same account authenticating to many systems in short time
RDP Lateral Movement
- Event 4624 with LogonType 10 (Remote Interactive)
- Chain of RDP sessions: A to B to C
- RDP from servers (servers should rarely initiate RDP)
- RDP to systems user has never accessed
SMB Lateral Movement
- Event 4624 with LogonType 3 (Network)
- Admin share access (C$, ADMIN$) from workstations
- Unusual account accessing file shares
- Service account authenticating interactively
WMI and PowerShell Remoting
- Event 4624 with LogonType 3 followed by process creation
- WinRM authentication events
- Remote PowerShell session establishment
Hunting Queries
Multiple Hosts Same Account (Lateral Movement)
Splunk:
index=wineventlog EventCode=4624 LogonType IN (3, 10)
| stats dc(dest) as unique_hosts values(dest) as hosts by user
| where unique_hosts > 5
| sort -unique_hosts
Admin Share Access
Splunk:
index=wineventlog EventCode=5140 ShareName IN ("\\*\C$", "\\*\ADMIN$")
| stats count by src_ip, user, ShareName, dest
| sort -count
Service Account Interactive Logon
Splunk:
index=wineventlog EventCode=4624 LogonType IN (2, 10) user=svc_*
| table _time, user, src_ip, dest, LogonType
Failed Logons Followed by Success
Splunk:
index=wineventlog EventCode IN (4625, 4624)
| transaction user maxspan=10m
| search EventCode=4625 AND EventCode=4624
| table _time, user, src_ip, dest
Kerberos Specific Hunting
Kerberoasting
Attackers request service tickets to crack offline:
- Event 4769 for service accounts with RC4 encryption (0x17)
- Single user requesting tickets for many services
- Unusual account requesting service tickets
Golden Ticket
Forged TGT allows access to any resource:
- TGS requests without corresponding TGT request
- Tickets with unusual lifetimes
- Domain admin activity from unexpected sources
Validation and False Positives
- Service accounts: May legitimately touch many systems
- Administrators: IT staff may have broad access patterns
- Automated processes: Scheduled tasks and monitoring may create patterns
- VPN and proxy: May mask true source location
Build allowlists for known patterns and validate anomalies against IT records.
Escalation Guidance
Escalate to incident response when:
- Confirmed unauthorized access to systems
- Evidence of credential theft (mimikatz artifacts, LSASS access)
- Lateral movement pattern across multiple systems
- Privileged account compromise suspected
- Access to sensitive data or systems
References
- MITRE ATT&CK: Lateral Movement (TA0008)
- MITRE ATT&CK: Credential Access (TA0006)
- Microsoft Security Auditing documentation
- CISA: Detecting Lateral Movement
- SANS: Spotting the Adversary with Windows Event Log Monitoring
Was this helpful?