Azure Sentinel for SIEM - Step-by-Step Execution

Process Flow - Azure Sentinel for SIEM

Start: Connect Data Sources

↓

Collect Security Data

↓

Analyze with AI & Rules

↓

Detect Threats

↓

Investigate Alerts

↓

Respond with Automation

↓

End

Azure Sentinel collects security data, analyzes it to detect threats, helps investigate alerts, and automates responses.

Execution Sample

Azure

1. Connect data sources
2. Collect logs and events
3. Apply detection rules
4. Generate alerts
5. Investigate and respond

This sequence shows how Azure Sentinel processes security data step-by-step.

Process Table

Step	Action	Input	Output	Notes
1	Connect Data Sources	Cloud apps, servers, devices	Data connectors active	Ready to receive data
2	Collect Security Data	Logs and events from sources	Raw security data stored	Continuous data flow
3	Apply Detection Rules	Raw data	Alerts generated	Rules and AI detect threats
4	Generate Alerts	Detected suspicious activity	Alert list updated	Alerts ready for review
5	Investigate Alerts	Alert details	Investigation results	Analyst reviews context
6	Respond with Automation	Investigation outcome	Automated actions triggered	Playbooks run automatically
7	End	All alerts processed	Security posture improved	Cycle repeats continuously

💡 All alerts processed and responses triggered; system waits for new data.

Status Tracker

Variable	Start	After Step 1	After Step 2	After Step 3	After Step 4	After Step 5	After Step 6	Final
Data Connectors	Inactive	Active	Active	Active	Active	Active	Active	Active
Security Data	None	None	Collected	Collected	Collected	Collected	Collected	Collected
Alerts	None	None	None	Generated	Updated	Reviewed	Responded	Processed
Investigation Status	None	None	None	None	Pending	Completed	Completed	Completed
Automation Status	None	None	None	None	None	Triggered	Triggered	Triggered

Key Moments - 3 Insights

Why do we need to connect data sources first before collecting data?

Because without active data connectors (see Step 1 in execution_table), Azure Sentinel cannot receive any security data to analyze.

What happens if detection rules do not generate any alerts?

If no alerts are generated (Step 3), then there is no suspicious activity detected, so investigation and response steps do not run.

How does automation improve the response process?

Automation triggers playbooks automatically after investigation (Step 6), speeding up response and reducing manual work.

Visual Quiz - 3 Questions

Look at the execution_table, what is the output after Step 3?

A Raw security data stored

B Data connectors active

C Alerts generated

D Automated actions triggered

💡 Hint

Check the 'Output' column for Step 3 in the execution_table.

At which step does Azure Sentinel start automated responses?

A Step 6

B Step 4

C Step 5

D Step 7

💡 Hint

Look for 'Automated actions triggered' in the 'Output' column.

If data connectors remain inactive, what will happen to security data collection?

A Data will still be collected

B No data will be collected

C Alerts will be generated anyway

D Automation will trigger without data

💡 Hint

Refer to the 'Data Connectors' variable in variable_tracker after Step 1.

Concept Snapshot

Azure Sentinel is a cloud SIEM tool.
Connect data sources first to collect security logs.
Apply detection rules to find threats.
Investigate alerts and respond automatically.
Automate with playbooks to speed up security actions.

Full Transcript

Azure Sentinel works by first connecting to your cloud apps, servers, and devices to collect security data. Once connected, it continuously gathers logs and events. Then, it applies detection rules and AI to find suspicious activities and generate alerts. Security analysts investigate these alerts to understand the threats. Finally, automated playbooks respond to threats quickly, improving your security posture. This cycle repeats continuously to keep your environment safe.

Practice

1. What is the main purpose of Azure Sentinel in security management?

easy

A. To provide cloud storage for application data

B. To collect and analyze security data for threat detection

C. To manage user passwords and authentication

D. To store backups of all user files

Solution

1. Step 1: Understand Azure Sentinel's role

   Azure Sentinel is designed to collect security data from various sources to detect threats.

2. Step 2: Compare options with Sentinel's function

   Only To collect and analyze security data for threat detection describes collecting and analyzing security data for threat detection, which matches Sentinel's purpose.

3. Final Answer:

   To collect and analyze security data for threat detection -> Option B

4. Quick Check:

   Azure Sentinel = threat detection [OK]

Hint: Remember: Sentinel = security data + threat detection [OK]

Common Mistakes:

• Confusing Sentinel with backup or storage services
• Thinking Sentinel manages passwords directly
• Assuming Sentinel is just cloud storage

2. Which of the following is the correct way to create an alert rule query in Azure Sentinel using Kusto Query Language (KQL)?

easy

A. GET SecurityEvent WHERE EventID = 4625

B. SELECT * FROM SecurityEvent WHERE EventID = 4625

C. FIND SecurityEvent WITH EventID 4625

D. SecurityEvent | where EventID == 4625

Solution

1. Step 1: Identify the query language used in Azure Sentinel

   Azure Sentinel uses Kusto Query Language (KQL), which uses pipe operators and 'where' clauses.

2. Step 2: Match the syntax to KQL

   SecurityEvent | where EventID == 4625 uses KQL syntax correctly: table name, pipe, and 'where' condition. Other options use SQL or invalid syntax.

3. Final Answer:

   SecurityEvent | where EventID == 4625 -> Option D

4. Quick Check:

   KQL uses pipes and 'where' [OK]

Hint: KQL uses pipes (|) and 'where' for filters [OK]

Common Mistakes:

• Using SQL syntax instead of KQL
• Missing pipe operator in query
• Using incorrect keywords like GET or FIND

3. Given the following KQL query in Azure Sentinel alert rule:
SecurityEvent | where EventID == 4625 | summarize count() by Account
What does this query output?

medium

A. A count of all events without grouping

B. A list of all successful login events

C. A count of failed login attempts grouped by user account

D. A list of accounts with no login attempts

Solution

1. Step 1: Analyze the query filters and aggregation

   The query filters SecurityEvent for EventID 4625, which means failed login attempts, then counts them grouped by Account.

2. Step 2: Understand the summarize clause

   'summarize count() by Account' groups results by Account and counts events per account.

3. Final Answer:

   A count of failed login attempts grouped by user account -> Option C

4. Quick Check:

   EventID 4625 = failed logins, grouped count = A count of failed login attempts grouped by user account [OK]

Hint: EventID 4625 means failed login; summarize groups counts [OK]

Common Mistakes:

• Confusing EventID 4625 with successful logins
• Ignoring the grouping by Account
• Thinking it lists accounts without attempts

4. You wrote this KQL alert rule query in Azure Sentinel:
SecurityEvent | where EventID = 4625 | summarize count() by Account
Why does this query fail to run correctly?

medium

A. Because the equality operator should be '==' not '=' in KQL

B. Because 'summarize' cannot be used with 'count()'

C. Because 'Account' is not a valid field in SecurityEvent

D. Because 'where' clause must come after 'summarize'

Solution

1. Step 1: Check the operator syntax in the 'where' clause

   KQL requires '==' for equality comparison, not a single '=' which is assignment in some languages.

2. Step 2: Validate other parts of the query

   'summarize count() by Account' is valid, and 'Account' is a common field. 'where' must come before 'summarize'.

3. Final Answer:

   Because the equality operator should be '==' not '=' in KQL -> Option A

4. Quick Check:

   KQL equality uses '==' not '=' [OK]

Hint: Use '==' for equality in KQL, not '=' [OK]

Common Mistakes:

• Using single '=' instead of '==' in KQL
• Misplacing 'where' after 'summarize'
• Assuming 'count()' is invalid with 'summarize'

5. You want to create an Azure Sentinel alert that triggers when there are more than 5 failed login attempts from the same account within 10 minutes. Which KQL query correctly implements this logic?

hard

A. SecurityEvent | where EventID == 4625 | where TimeGenerated > ago(10m) | summarize FailedAttempts = count() by Account | where FailedAttempts > 5

B. SecurityEvent | where EventID == 4625 | summarize count() by Account | where count_ > 5

C. SecurityEvent | where EventID == 4625 and TimeGenerated < ago(10m) | summarize count() by Account | where count_ > 5

D. SecurityEvent | where EventID == 4625 | summarize count() by Account, TimeGenerated | where count_ > 5

Solution

1. Step 1: Filter failed login events within last 10 minutes

   SecurityEvent | where EventID == 4625 | where TimeGenerated > ago(10m) | summarize FailedAttempts = count() by Account | where FailedAttempts > 5 uses 'where TimeGenerated > ago(10m)' to filter recent events correctly.

2. Step 2: Group by Account and count attempts, then filter counts over 5

   SecurityEvent | where EventID == 4625 | where TimeGenerated > ago(10m) | summarize FailedAttempts = count() by Account | where FailedAttempts > 5 summarizes counts by Account and filters where count > 5, matching the requirement.

3. Final Answer:

   SecurityEvent | where EventID == 4625 | where TimeGenerated > ago(10m) | summarize FailedAttempts = count() by Account | where FailedAttempts > 5 -> Option A

4. Quick Check:

   Filter by time + count > 5 per account = SecurityEvent | where EventID == 4625 | where TimeGenerated > ago(10m) | summarize FailedAttempts = count() by Account | where FailedAttempts > 5 [OK]

Hint: Filter time first, then count and filter by count > 5 [OK]

Common Mistakes:

• Not filtering events by time range
• Using incorrect logical operators in filters
• Grouping by TimeGenerated causing wrong counts