Microsoft Sentinel — Data Hierarchy

How Data Flows Through the Sentinel Stack

Click any layer to expand
01
LOG SOURCES
Where all data originates — every event, login, and alert
☁️
Azure Resources
VMs, NSGs, Key Vault, Storage, Azure AD activity
📧
Microsoft 365
Exchange, Teams, SharePoint, OneDrive, Defender for O365
🛡️
Microsoft Defender
Defender for Endpoint, Identity, Cloud Apps
🖥️
On-Premises
Windows / Linux servers, Active Directory, DNS
🌐
Network Devices
Firewalls, routers, switches — Cisco, Palo Alto, Fortinet
🔗
Third-Party / Cloud
AWS, GCP, SaaS apps, custom REST API sources
Key point: Sentinel is source-agnostic — it can ingest from virtually any platform as long as data reaches Log Analytics Workspace in a supported format.
raw logs and events
02
DATA CONNECTORS
The bridge — normalises and forwards data into Sentinel
🔌
Built-in Connectors
1-click connectors for all Microsoft services and 200+ partners
📡
CEF / Syslog
Common Event Format via Log Analytics Agent for network devices and Linux
🤖
AMA / MMA Agent
Azure Monitor Agent installed on VMs and on-prem servers
🔄
REST API / DCR
Data Collection Rules and custom log ingestion API for any source
Logic App Connector
Custom ingestion pipelines using Azure Logic Apps for any REST source
Key point: Connectors normalise data into a common schema before it enters LAW. This is why one KQL query can work across multiple source types.
normalised and ingested
03
LOG ANALYTICS WORKSPACE (LAW)
The data store — all logs live here, queried via KQL
📋
Log Tables
SecurityEvent, SigninLogs, OfficeActivity, CommonSecurityLog, Syslog...
🔍
KQL Engine
Kusto Query Language — search, filter, join, aggregate across all tables
🗂️
Data Retention
Hot tier (90 days default), Archive tier (up to 7 years)
📊
Workbooks
Interactive dashboards built directly on top of LAW query results
TableWhat It ContainsCommon Use
SigninLogsAzure AD authentication eventsImpossible travel, brute force, MFA analysis
SecurityEventWindows security event IDsLogon events 4624/4625, privilege use
OfficeActivityM365 — email, SharePoint, TeamsMail forwarding rules, file downloads
AzureActivityAzure resource changesWho changed what in your subscription
CommonSecurityLogCEF-format logs from firewallsNetwork threat detection
ThreatIntelligenceIndicatorIOCs — IPs, domains, hashesMatching logs against known threats
Splunk equivalent: LAW is Splunk's index + search head combined. KQL is your SPL. Tables are your sourcetypes.
Sentinel reads and analyses LAW data
04
MICROSOFT SENTINEL — SIEM + SOAR
The brain — detects threats, enriches context, raises incidents
⚠️
Analytics Rules
Scheduled, NRT, Fusion, ML-based — run KQL on schedule, create alerts
🚨
Incidents
Grouped alerts with investigation graph, entities, timeline, severity
📌
Watchlists
CSV reference lists — VPN IPs, VIP users, known-good assets
🕵️
Threat Intelligence
IOC feeds matched against all ingested logs in real time
👤
UEBA
User and Entity Behaviour Analytics — baselines normal, flags deviations
🏹
Hunting Queries
Proactive threat hunting using KQL across all historical data
📓
Notebooks
Jupyter + Python for advanced investigation, ML, geolocation analysis
Internal flow: Analytics Rule fires → Alert created → Alerts grouped → Incident raised → Entities extracted (user, IP, host, URL) → Assigned to analyst → Investigation begins
incident triggers response
05
RESPONSE AND AUTOMATION
What happens after a threat is confirmed
🎭
Playbooks
Logic Apps triggered on incidents — automated response workflows
🤖
Automation Rules
Auto-assign, triage, tag incidents, or trigger a playbook on creation
🔧
SOAR Actions
Block IP, disable user, isolate device, send Teams and email alerts
Example: Incident created → Automation Rule fires → Playbook runs → Azure AD account disabled + SOC notified via Teams — all within seconds, no human needed.
outputs delivered to
06
OUTPUTS
Where the results of SOC work land
🖥️
SOC Dashboard
Real-time incident queue, analyst workload, alert trends
🔔
Alerts and Emails
Automated notifications to analysts, managers, on-call teams
🎫
ITSM / ServiceNow
Incidents pushed to ticketing systems for tracking and SLA management
📈
Reports and Metrics
MTTD, MTTR, false positive rate, coverage dashboards for leadership
LAYERS:
Log Sources
Data Connectors
Log Analytics Workspace
Microsoft Sentinel
Response and Automation
Outputs