High School IntermediateModule I4Lesson 6 of 8

I4.6 Normal vs Suspicious Patterns

Learn how defenders compare fictional activity with baselines, ownership, schedules, roles, applications, destinations, privilege, change records, and multi-source evidence before deciding what deserves escalation.

Lesson Progress

Normal vs Suspicious Patterns

High School IntermediateI4: Logs and Event Monitoring • Lesson 6 of 8

75% complete

Readiness Check

Before You Start

0/5 ready

Professional Hook

Unusual Does Not Automatically Mean Malicious

A late-night sign-in, large transfer, rare destination, repeated error, new process, high-severity alert, or denied connection may be normal in one environment and concerning in another. Defenders compare the pattern with role, ownership, purpose, schedule, privilege, baseline, and corroborating evidence.

Weak response

“It has never happened before, so it must be malicious.”

Strong response

“Measure the deviation, identify the baseline, add role and owner context, correlate related sources, and state the classification with confidence and limitations.”

Objective 1

Explain the difference between normal, expected, unusual, suspicious, administrative, failed, and evidence-incomplete activity.

Objective 2

Use baselines, asset role, user role, schedule, frequency, volume, source, destination, privilege, change, and owner context to evaluate fictional patterns.

Objective 3

Recognize why rare activity, high severity, denied actions, failed sign-ins, large transfers, and unusual times do not automatically prove malicious intent.

Objective 4

Prioritize fictional patterns by evidence strength, potential impact, confidence, urgency, and safe next action.

Objective 5

Create a professional fictional Pattern Analysis Matrix that separates confirmed facts, reasonable conclusions, alternate explanations, evidence gaps, and escalation decisions.

Why This Matters

Good Pattern Analysis Reduces Both Missed Risk and Unnecessary Escalation

Weak analysis can ignore important combinations of signals or overwhelm teams with false alarms. Strong analysis uses updated baselines, multi-source evidence, context, confidence, impact, and safe prioritization.

Pattern Dimensions

Ten Questions That Change Pattern Meaning

Time

Did the activity occur during normal hours, an approved maintenance window, travel, backup, deadline work, or scheduled automation?

Unusual time can increase review priority, but it does not prove malicious intent.

Frequency

Did the event happen once, repeatedly, on a fixed schedule, in a burst, or across many accounts and devices?

Repetition may reflect retries, stale credentials, health checks, monitoring, or a real issue.

Volume

Did the amount of data, requests, errors, sign-ins, changes, or alerts differ significantly from the approved baseline?

Large activity can be expected during backups, updates, reporting, migration, or class-wide use.

Source

Did the activity come from an expected device, user, process, service, network, VPN, gateway, or management system?

Shared infrastructure, proxies, mobile networks, and translation can change the apparent source.

Destination

Did the activity reach an approved application, server, cloud service, vendor, update source, or internal zone?

A new destination may still be approved, and a familiar destination does not guarantee safe purpose.

Privilege

Was the activity performed by a standard user, administrator, service identity, temporary role, or highly privileged account?

Privileged activity requires stronger ownership and approval context, even when it succeeds.

Result

Did the action succeed, fail, get denied, get blocked, time out, retry, or partially complete?

Success does not automatically mean approved, and failure does not automatically mean malicious.

Change context

Was there an approved deployment, password reset, update, rule change, test, migration, restart, or support action?

Documentation can be incomplete or different from the technical sequence, so it still requires validation.

Asset role

Is the device a workstation, server, backup system, management platform, kiosk, lab machine, or public-facing service?

The same behavior may be normal on one asset and inappropriate on another.

Historical pattern

Has the same user, application, process, destination, or schedule appeared before under approved conditions?

A known pattern can still become risky after ownership, purpose, or configuration changes.

Classification Model

Seven Defensive Classification Labels

Normal

Meaning

Matches an established baseline with expected users, devices, applications, timing, destinations, volume, and results.

Example

A backup server transfers its usual nightly volume to the approved storage destination during the documented window.

Strong action

Document the match and continue routine monitoring.

Expected

Meaning

May be new or rare but is supported by a valid owner, business purpose, change record, schedule, or approved workflow.

Example

A new vendor destination appears during an approved application migration with owner and validation evidence.

Strong action

Confirm the approved context and monitor the new baseline.

Unusual

Meaning

Differs from the baseline but currently lacks enough evidence to call suspicious.

Example

A user signs in two hours later than usual from the same managed device and approved application.

Strong action

Add user, device, schedule, application, and owner context before deciding.

Suspicious

Meaning

Contains multiple concerning signals or corroborating evidence that justify focused defensive review.

Example

A privileged account authenticates from a new device, receives denied MFA prompts, and creates an unexpected session outside the approved schedule.

Strong action

Preserve evidence, verify through approved channels, and escalate according to policy.

Administrative

Meaning

Represents approved management, maintenance, support, monitoring, configuration, deployment, testing, or recovery activity.

Example

A management server connects to many internal systems during a documented patching window.

Strong action

Confirm ownership, scope, schedule, results, and change closure.

Failed

Meaning

An attempted action did not complete successfully and requires cause and impact context.

Example

A retired application repeatedly attempts a denied connection to an old server address.

Strong action

Identify the source process and owner, correct the stale configuration, and validate recovery.

Evidence-incomplete

Meaning

The available records are insufficient to classify the pattern responsibly.

Example

A DNS request appears, but endpoint, firewall, proxy, application, and owner records are missing.

Strong action

Preserve the confirmed event, identify the gaps, and request authorized supporting evidence.

Core Concept

Patterns Become Meaningful When Multiple Dimensions Agree

Time, frequency, volume, source, destination, privilege, result, asset role, history, and change context each provide part of the picture. The strongest classifications explain how those parts connect and what evidence remains missing.

Measure

What exactly changed from the approved baseline?

Contextualize

Which user, device, application, owner, schedule, and purpose apply?

Correlate

Which independent sources agree or conflict?

Classify

Is the activity normal, expected, unusual, suspicious, administrative, failed, or incomplete?

Act

What is the safest authorized next step, and how will it be validated?

Signal Combinations

One Signal Is Weak; Combined Context Is Stronger

One unusual sign-in outside normal hours

Lower-risk context

Expected device, expected application, approved travel, successful MFA, no sensitive activity, and owner confirmation.

Higher-risk context

New device, privileged account, repeated failures, denied MFA, sensitive application, and no owner explanation.

Defensive lesson

Time alone is weak; combined identity, device, privilege, MFA, and application context changes priority.

Large outbound transfer

Lower-risk context

Approved backup, report export, migration, cloud synchronization, known process, expected destination, and change ticket.

Higher-risk context

Unknown process, sensitive source, unfamiliar destination, unusual time, no owner, and policy bypass.

Defensive lesson

Volume becomes more meaningful when process, data role, destination, owner, and change evidence agree.

Repeated failed sign-ins

Lower-risk context

Same expected device, stale saved credential, recent password reset, application retry, and failures stop after update.

Higher-risk context

Many accounts, many sources, unfamiliar devices, lockouts, later success elsewhere, and failed MFA.

Defensive lesson

Failure count alone is not enough; distribution, device, application, reset, and later-success context matter.

One device connects to many internal systems

Lower-risk context

Approved inventory, monitoring, patching, backup, or management server during a documented window.

Higher-risk context

User workstation, unknown process, unusual ports, sensitive zones, no management role, and failed authentication.

Defensive lesson

Asset role and process ownership can completely change how the pattern is interpreted.

High-severity endpoint alert

Lower-risk context

Known test file, approved administrative tool, expected lab activity, completed remediation, and no related events.

Higher-risk context

Successful execution, new persistence, broad exclusion, privileged account, network activity, and multiple devices.

Defensive lesson

Severity is one clue; process, action, scope, remediation, persistence, and correlation determine urgency.

Many web errors

Lower-risk context

Old bookmark, expired session, application deployment, maintenance window, or temporary back-end restart.

Higher-risk context

Sensitive path, unusual method, many sources, growing rate, no maintenance, and repeated successful access attempts.

Defensive lesson

Error counts need path, method, source, application, maintenance, and response-sequence context.

New scheduled task or startup item

Lower-risk context

Approved software deployment, known publisher, expected path, owner, change record, and successful validation.

Higher-risk context

Unknown publisher, user-writable path, privileged account, no owner, repeated network activity, and no change record.

Defensive lesson

Newness becomes more concerning when path, publisher, privilege, owner, and related activity are also unusual.

Missing logs from one device

Lower-risk context

Device powered off, traveling, sleeping, offline, agent upgrade, or documented collector maintenance.

Higher-risk context

Device reported online, sensitive role, repeated collection failure, local logs unavailable, and protection state unknown.

Defensive lesson

Missing telemetry confirms a visibility gap; device state and collection evidence shape urgency.

Evidence Matrix

What Pattern Evidence Can and Cannot Prove

Evidence source

Baseline record

Can support

Expected users, devices, schedules, applications, destinations, volume, frequency, and normal results.

Limitation

Baselines can be stale, incomplete, seasonal, or too broad.

Evidence source

Asset inventory

Can support

Device role, owner, operating system, exposure, sensitivity, management status, and approved software.

Limitation

Inventory can lag behind actual configuration or ownership changes.

Evidence source

Identity context

Can support

Account role, privilege, groups, MFA, expected devices, schedule, lifecycle, and owner.

Limitation

Account context does not prove the physical person or every action.

Evidence source

Technical event records

Can support

Observed time, source, destination, process, action, result, severity, volume, and sequence.

Limitation

Technical events may not explain purpose, authorization, or business impact.

Evidence source

Change record

Can support

Approved owner, reason, scope, schedule, test, rollback, expected behavior, and validation.

Limitation

The documented change may not match the exact technical activity.

Evidence source

User or owner report

Can support

Expected activity, observed impact, travel, device use, business need, and timing.

Limitation

Reports can be incomplete, delayed, approximate, or mistaken.

Evidence source

Historical comparison

Can support

Whether the same user, device, process, destination, volume, or schedule appeared before.

Limitation

Past normality does not guarantee current safety or approval.

Evidence source

Correlated multi-source evidence

Can support

Agreement across authentication, endpoint, system, application, network, web, and change records.

Limitation

Correlation quality depends on time accuracy, field consistency, retention, and source coverage.

Prioritization

Eight Factors That Shape Review Priority

Evidence strength

How many independent fictional sources support the pattern, and how direct are they?

More independent, well-aligned evidence usually increases confidence.

Potential impact

Could the activity affect a sensitive account, critical service, large user group, protected data set, or important business function?

Higher potential impact can justify faster review even when confidence is moderate.

Privilege

Does the pattern involve an administrator, service identity, security control, management platform, or broad-access role?

Privileged context increases the importance of ownership and authorization evidence.

Scope

Does the pattern affect one event, one user, one device, many systems, many accounts, or multiple network zones?

Broader scope often raises urgency and coordination needs.

Persistence

Did the pattern happen once, stop after correction, repeat regularly, or continue despite remediation?

Continued or recurring activity usually deserves higher priority.

Control outcome

Was the action allowed, denied, blocked, quarantined, contained, retried, or successful?

Completed activity may require more impact review, but blocked attempts can still reveal important patterns.

Context quality

Are owner, baseline, change, schedule, device, application, destination, and user details available?

Weak context lowers confidence and may require evidence collection before classification.

Time sensitivity

Is the activity ongoing, recent, recurring, or tied to an active user, service, or business disruption?

Ongoing activity may require immediate preservation and escalation.

Defensive Workflow

Classify Patterns in Six Steps

1

Define the pattern

State exactly what changed in time, frequency, volume, source, destination, privilege, result, or behavior.

2

Identify the baseline

Find the approved users, devices, applications, schedules, destinations, volumes, and normal results.

3

Add context

Compare asset role, identity role, owner, change, travel, maintenance, application purpose, and business need.

4

Correlate sources

Connect authentication, endpoint, system, application, network, web, ticket, owner, and monitoring evidence.

5

Classify and prioritize

Choose normal, expected, unusual, suspicious, administrative, failed, or evidence-incomplete and state confidence.

6

Respond and improve

Preserve evidence, assign an owner, take the safest authorized action, validate the result, and update the baseline.

Example Findings

Classification Should Be Evidence-Based and Actionable

Finding

Late-night sign-in from an expected device

Supporting facts

Same managed laptop, approved application, successful MFA, no sensitive changes, and user confirms deadline work.

Classification

Expected

Confidence

High

Strong action

Document owner confirmation and include the approved late-work pattern in the baseline review.

Finding

Repeated failed sign-ins after password reset

Supporting facts

Same device, mail client reports stale credential, failures stop after update, and later MFA succeeds.

Classification

Expected failure pattern

Confidence

High

Strong action

Document stale credential, validate recovery, and monitor for recurrence.

Finding

Rare DNS request without endpoint or connection evidence

Supporting facts

One query appears, but process, firewall, proxy, application, and owner records are unavailable.

Classification

Evidence-incomplete

Confidence

High

Strong action

Preserve the DNS event and request authorized supporting evidence before deciding.

Finding

Privileged sign-in from new device with denied MFA prompts

Supporting facts

New device, unusual time, privileged account, two denied prompts, no travel record, and no owner confirmation.

Classification

Suspicious

Confidence

Medium

Strong action

Preserve evidence, verify through approved channels, review sessions and device state, and escalate promptly.

Finding

Management server connects to many endpoints

Supporting facts

Known management asset, approved patching window, expected ports, named owner, and successful patch validation.

Classification

Administrative

Confidence

High

Strong action

Confirm scope and change closure, then continue routine monitoring.

Finding

Large approved report upload

Supporting facts

Known process, expected destination, approved change, matching volume, successful application result, and owner validation.

Classification

Expected

Confidence

High

Strong action

Document the new approved destination and continue monitoring.

Key Vocabulary

Pattern Analysis Terms

Baseline

An approved description of expected users, devices, applications, destinations, schedules, volumes, and behaviors.

Normal

Activity that matches an established and approved pattern for the system, user, application, or business process.

Expected

Activity that is justified by role, owner, schedule, change, ticket, workflow, or documented purpose.

Unusual

Activity that differs from the known baseline but is not automatically harmful.

Suspicious

Activity with enough concerning context or corroboration to justify focused defensive review.

Administrative

Authorized management, maintenance, support, configuration, deployment, monitoring, or recovery activity.

Failed

An attempted action that did not complete successfully.

Evidence-incomplete

Activity that cannot be classified confidently because important records or context are missing.

Frequency

How often an event or pattern occurs within a defined period.

Volume

How much activity, data, traffic, or change occurred within a defined period.

Deviation

A measurable difference from an approved baseline.

False positive

An alert or pattern that appears concerning but is explained by expected or benign activity after review.

Fake Dashboard

Fake Pattern Analysis Dashboard

Training dashboard for the fictional Northstar Learning Services monitoring environment.

Expected patterns

31

Validated backups, updates, patching, report exports, travel, and scheduled support activity.

Unusual patterns

9

Five have reasonable owner explanations, while four still require supporting evidence.

Suspicious patterns

3

All involve combined identity, device, privilege, MFA, or persistence signals and have been escalated.

Fake SOC Alert

Privileged Sign-In from a New Device with Denied MFA Prompts

Source: Fake Pattern Correlation Platform • Time: 07:42 PM

High Severity
A fictional privileged account signs in outside its normal schedule from a new unmanaged device. Two MFA prompts are denied, no approved travel or support record exists, and a later session attempt targets a sensitive administrative portal.
Defensive recommendation: Preserve identity, MFA, device, session, application, network, owner, and change evidence; verify through an approved channel; review active sessions and account changes; assign the incident owner; and follow the authorized escalation process.

Fake Log Panel

Fake Pattern Evidence

training-log-viewer.log
19:40:08 AUTH account='training-admin-6' device='unknown-device-91' result='failed' reason='mfa_required'
19:40:16 MFA account='training-admin-6' result='denied' source='unknown-device-91'
19:40:29 AUTH account='training-admin-6' device='unknown-device-91' result='failed' reason='mfa_denied'
19:41:03 MFA account='training-admin-6' result='denied' source='unknown-device-91'
19:41:45 SESSION app='admin-portal' result='not_created'
19:42:00 BASELINE account='training-admin-6' normal_hours='08:00-17:00' managed_devices='2'
19:42:05 INVENTORY device='unknown-device-91' status='not_found'
19:42:10 CHANGE approved_travel='none' support_ticket='none'
19:42:18 OWNER confirmation='not_yet_available'
19:42:30 CLASSIFICATION label='suspicious' confidence='medium' escalation='required'

Training note: this is fake data for defensive analysis practice only.

Analyze the Evidence

Which Pattern Classification Is Best Supported?

The account is privileged.
The device is new and not present in managed inventory.
The activity occurs outside the account's normal schedule.
Two MFA prompts are denied.
No successful session is created.
No approved travel, support, or change record exists.
Owner confirmation is not yet available.
The attempted application is a sensitive administrative portal.

Which classification is strongest?

Common Mistakes

Mistakes That Weaken Pattern Analysis

Calling every unusual event suspicious.
Calling every high-severity alert a confirmed incident.
Treating rare activity as automatically malicious.
Treating blocked or denied actions as proof of compromise.
Treating successful actions as automatically approved.
Using only one field such as time, source, destination, or volume.
Ignoring asset role, user role, application purpose, owner, schedule, and change context.
Comparing activity with a stale or overly broad baseline.
Ignoring seasonal, maintenance, travel, deadline, backup, update, and migration patterns.
Assigning exact confidence when evidence is incomplete or conflicting.
Escalating without preserving the original evidence and documenting the reason.
Publishing real user behavior, addresses, destinations, timelines, account details, or internal baselines.

Safe Practice Lab

Classify a Fictional Pattern Set

Fictional Environment

Meadowbrook Pattern Review

Review sixteen fictional patterns involving authentication, endpoint, system, application, network, web, change, support, inventory, and baseline evidence.

Required Analysis

  1. State the exact deviation in time, frequency, volume, source, destination, privilege, or result.
  2. Identify the relevant baseline and whether it is current.
  3. Add asset, identity, application, owner, schedule, destination, and change context.
  4. Correlate at least two independent evidence sources.
  5. Classify each pattern using one of the seven labels.
  6. Assign confidence, impact, urgency, owner, and safe next action.
  7. Explain what evidence would change the classification.
Use only supplied fictional evidence. Do not monitor real people, inspect private activity, access accounts, test systems, alter alerts, or publish real user behavior, destinations, schedules, devices, account details, or internal baselines.

Scenario Decision Lab

A Large Transfer Occurs During an Approved Backup Window

A fictional server transfers three times its usual volume to the approved backup destination. The backup owner confirms a full monthly archive, but file-level validation is still pending.

Scenario Decision Lab

A New Startup Item Appears After Approved Software Deployment

A fictional managed workstation gains a new startup item from the approved application folder. The publisher, path, owner, and change record match the deployment, and no unusual network or account activity follows.

Defender Habits

Normal vs Suspicious Patterns Checklist

Check Your Understanding

I4.6 Mini Quiz: Normal vs Suspicious Patterns

Choose your answers first. Explanations appear only after submission.

1. What makes unusual activity different from suspicious activity?

2. Why is a baseline important?

3. What does a rare DNS request directly prove?

4. Which pattern deserves the highest review priority?

5. What should happen when evidence is incomplete?

6. Why can a large outbound transfer be expected?

7. What should happen after an expected new pattern is validated?

Portfolio Prompt

Portfolio Prompt

Create a fictional Pattern Analysis Matrix containing at least sixteen patterns across authentication, endpoint, system, application, network, web, inventory, baseline, change, and support evidence. For each pattern, include deviation, baseline, asset role, identity role, application, time, frequency, volume, source, destination, privilege, result, owner, change context, corroborating sources, classification, confidence, potential impact, urgency, evidence gaps, safe next action, and validation plan.

Use only fictional users, devices, applications, destinations, schedules, alerts, baselines, and organizations.
Include at least two normal, two expected, two unusual, two suspicious, two administrative, two failed, and two evidence-incomplete patterns.
Explain what evidence would change each classification.
Do not include real monitoring exports, user activity, addresses, domains, device names, account details, schedules, or internal baselines.

Key Takeaways

What You Should Remember

1.Unusual activity differs from a baseline but is not automatically suspicious.
2.Strong pattern analysis combines time, frequency, volume, source, destination, privilege, result, role, history, and change context.
3.Baselines must be current, specific, and connected to owners and business purpose.
4.Rare, severe, denied, failed, or high-volume activity still requires corroboration and evidence limits.
5.Evidence-incomplete is a valid classification when important context is missing.
6.Professional prioritization considers evidence strength, impact, privilege, scope, persistence, control outcome, context quality, and time sensitivity.

Navigation

Continue Module I4