Investigating Entra ID Sign-In Alerts: Logs, Pivots, and Containment
· 44:38 · Watch on YouTube ↗
Summary
A practical guide to investigating Entra ID alerts by correlating sign-in, audit, and Graph activity logs, identifying suspicious token use and persistence, and containing compromised accounts.
Key points
- Review interactive and non-interactive sign-ins, audit records, Microsoft 365 activity, and Graph activity logs; missing log collection can leave investigators unable to determine what an attacker did.
- Correlate events using session and token identifiers where available. Compare IPs, user agents, authentication flows, and activity against each user's normal baseline.
- Treat patterns such as repeated MFA prompts followed by success, unusual device-code sign-ins, and stolen-token use from a new IP as investigation leads.
- After a suspicious sign-in, look for persistence and follow-on activity, including new authenticators, app credentials, inbox rules, forwarding, mailbox access, and file sharing.
- Search across users and related email recipients to assess whether the activity affected other accounts or exposed partners and customers to follow-on phishing or business email compromise.
- Containment may require blocking access, revoking sessions, removing attacker-added authenticators, app secrets, consent, and mail rules, then continuing to monitor for activity.
- Preventive measures discussed include blocking device-code flow where unnecessary, using phishing-resistant MFA, restricting user consent, protecting privileged roles, and enabling log collection before an incident.
Article Details
- Topics
- [00:42] Investigating Entra ID identity alerts and choosing relevant logs
- [01:51] Sign-in, audit, and Graph activity logs and common collection gaps
- [03:53] Investigation fields, authentication failures, and risk signals
- [07:17] Correlating audit events with sign-in sessions and pivoting across activity
- [10:00] Baselines, investigation workflow, and assessing attack scope
- [15:11] Attack scenarios including device-code phishing, token use, and mailbox compromise
- [27:02] Containment, detection engineering, and preventive controls
MITRE ATT&CK
T1098.001 · Additional Cloud CredentialsThe attacker adds a client secret to an app registration to retain access as the app, even after the user's password and sessions are addressed.T1528 · Steal Application Access TokenIn the device-code phishing scenario, the victim completes authentication and the attacker obtains a token for later use.T1564.008 · Email Hiding RulesThe presenter describes inbox rules that move or hide messages so attackers can continue mailbox activity without the user seeing replies.T1621 · Multi-Factor Authentication Request GenerationThe presenter describes MFA push fatigue: an attacker repeatedly prompts a user until one request is approved.
