Azure Sentinel for SIEM - Time & Space Complexity

Time Complexity: Azure Sentinel for SIEM

O(n)

Understanding Time Complexity

We want to understand how the time to analyze security data grows as more data is collected in Azure Sentinel.

Specifically, how does the number of operations change when processing more alerts and logs?

Scenario Under Consideration

Analyze the time complexity of querying and alerting in Azure Sentinel.

// Pseudo-azure code for Sentinel query and alert
let alerts = SecurityAlert
| where TimeGenerated > ago(1d)
| where Severity == 'High'
| summarize count() by AlertName

alerts
| where count_ > 10
| project AlertName, count_
    

This sequence queries high severity alerts from the last day, counts them by alert type, and filters for frequent alerts.

Identify Repeating Operations

Identify the API calls, resource provisioning, data transfers that repeat.

• Primary operation: Querying logs from the SecurityAlert table.
• How many times: Once per query, but the query scans all alerts in the time range.
• Data aggregation: Counting alerts by type involves scanning all matching records.

How Execution Grows With Input

As the number of alerts grows, the query scans more records, so the time grows roughly in proportion to the number of alerts.

Input Size (n alerts)	Approx. Api Calls/Operations
10	Scan 10 alerts
100	Scan 100 alerts
1000	Scan 1000 alerts

Pattern observation: The work grows linearly as the number of alerts increases.

Final Time Complexity

Time Complexity: O(n)

This means the time to process alerts grows directly with the number of alerts collected.

Common Mistake

[X] Wrong: "Querying alerts always takes the same time regardless of data size."

[OK] Correct: The query scans all matching alerts, so more alerts mean more work and longer time.

Interview Connect

Understanding how data size affects query time in Azure Sentinel helps you design efficient security monitoring solutions.

Self-Check

"What if we added indexing or partitioning to the SecurityAlert table? How would the time complexity change?"

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