Defensive Security

Blue Team
Introduction

Blue Team Fundamentals, Logs, SIEM Thinking, IOCs & an Introduction to MITRE ATT&CK

A. Yahiaoui, Security·Updated: August 15, 2026·Hands-on labs included

Use ← → or swipe to navigate

Guide Map

What We're Covering Today

01 · FundamentalsThe Blue Team mission, the roles, the mindset
02 · Logs & SIEMLogs, SIEM thinking, and IOCs, turning noise into signal
03 · MITRE ATT&CKA shared language for describing adversary behavior
04 · LabsTriage a log set, map an attack, write a detection rule
01

Blue Team Fundamentals

Who defenders are, what they own, and how they think

01 · Blue Team Fundamentals

What Is "Blue Team"?

The defensive side of cybersecurity, the people, process, and tooling that keep an organization's systems standing.

  • Owns detection, response, and hardening, not just firewalls and antivirus
  • Operates continuously: attackers don't work business hours
  • Judged by one question, how fast did we notice, and how fast did we act?
If Red Team asks "can we get in?", Blue Team asks "would we even know?"
01 · Blue Team Fundamentals

Red, Blue & Purple

Red Team

Offense. Simulates real adversaries under controlled rules of engagement to find gaps before attackers do.

Blue Team

Defense. Detects, responds to, and hardens against intrusions. Owns the SOC.

Purple Team

Not a separate team: a way of working. Red executes a technique, Blue tries to catch it, both sides compare notes and close the gap.

01 · Blue Team Fundamentals

The Mission, in Six Functions

Most Blue Team programs are built around the NIST Cybersecurity Framework (CSF 2.0), six functions covering the full lifecycle of managing risk.

GovernSet the strategy, policy, and risk appetite everything else runs on
IdentifyKnow your assets, data, and exposure
ProtectControls that reduce likelihood and impact
DetectSpot malicious activity while it's happening
RespondContain, eradicate, communicate
RecoverRestore operations and feed lessons back in

CSF 2.0 (2024) added Govern as the sixth function. This guide lives mostly in Detect, with a foot in Respond.

01 · Blue Team Fundamentals

Who's on the Team

Tier 1 AnalystFirst eyes on alerts, initial triage
Tier 2 AnalystInvestigates, correlates evidence, escalates
Tier 3 / Threat HunterProactively hunts for what alerts didn't catch
Incident ResponderContains and remediates active incidents
Detection EngineerBuilds and tunes the rules everyone else relies on

Today's labs walk the Tier 1 → Detection Engineer path: triage, then investigate, then build a rule.

01 · Blue Team Fundamentals

The Blue Team Mindset

Assume breachBuild as if someone is already inside
Defense in depthNo single control should be the only thing standing
Visibility firstYou can't defend what you can't see
Least privilegeShrink what any one compromise can reach
Document everythingToday's noise is tomorrow's evidence
01 · Blue Team Fundamentals

How Attacks Unfold: the Cyber Kill Chain

Lockheed Martin's model breaks an intrusion into stages. Every stage is a chance to break the chain.

ReconStudy the target
→
WeaponizationBuild the payload
→
DeliveryGet it to the target
→
ExploitationTrigger the payload
→
InstallationEstablish a foothold
→
C2Open a control channel
→
Actions on ObjectivesDo what they came for

Later today we'll meet MITRE ATT&CK, a far more detailed map of this same territory.

02

Logs, SIEM Thinking & IOCs

Turning raw noise into a signal worth acting on

02 · Logs, SIEM & IOCs

What Is a Log?

A timestamped record of an event, who did what, where, and when.

  • Generated constantly, by nearly everything: OS, apps, network gear, cloud services
  • On its own, one log line rarely means much
  • Blue Team's job is to find the lines that matter, and the pattern across many that don't
If it's not logged, it didn't happen, from a detection standpoint.
02 · Logs, SIEM & IOCs

Where Logs Come From

AuthenticationWho logged in, from where, and whether it workedWindows Security Event Log · /var/log/auth.log
Network / FirewallWhich connections were allowed or blocked
DNSWhat domains a host resolved, often the first sign of C2 or exfil
Web / ProxyHTTP requests, user-agents, response codes
EDR / AntivirusProcess, file, and registry activity on the endpoint
CloudWho did what in your cloud consoleCloudTrail · Azure Activity Log
02 · Logs, SIEM & IOCs

The Log Triage Pipeline

CollectPull logs from every source into one place
→
NormalizeForce different formats into a common shape
→
CorrelateLink related events across sources and time
→
AlertRaise a flag when a rule's conditions are met
→
InvestigateAn analyst decides, noise, or real?
→
Close / EscalateDocument the call, or hand it to IR
02 · Logs, SIEM & IOCs

Case File 01: Reading the Log

Log sample, web01 / sshd
109:14:02 sshd[2213]: Failed password for root from 185.221.104.7 port 51422 ssh2
209:14:05 sshd[2213]: Failed password for root from 185.221.104.7 port 51430 ssh2
309:14:09 sshd[2213]: Failed password for root from 185.221.104.7 port 51439 ssh2
409:14:13 sshd[2213]: Failed password for root from 185.221.104.7 port 51447 ssh2
509:14:18 sshd[2213]: Failed password for root from 185.221.104.7 port 51455 ssh2
609:14:47 sshd[2213]: Accepted password for root from 185.221.104.7 port 51502 ssh2
  1. 1Same source IP, every attempt, not a typo, a pattern
  2. 2Five failures in under twenty seconds, no human types that fast
  3. 3Then... a success. This is the line that should worry you most.

We'll come back to this exact log in Case File 02, and in tonight's labs.

02 · Logs, SIEM & IOCs

What Is a SIEM?

Security Information and Event Management, pulls logs from everywhere and gives you one place to search, correlate, and get alerted.

  • Aggregation, every log source, one index
  • Correlation, connect events that mean something together
  • Alerting, flag activity that matches a rule
  • Dashboards, see the state of your environment at a glance
A SIEM doesn't replace an analyst's judgment. It just makes sure the right haystack lands in front of them.
02 · Logs, SIEM & IOCs

The SIEM Landscape

SplunkIndustry standard; powerful search language (SPL)
Elastic StackElasticsearch, Logstash, Kibana, open-source-friendly
Microsoft SentinelCloud-native, tight Azure / Microsoft 365 integration
WazuhOpen-source, a strong choice for learning and home labs
Prometheus + Loki + GrafanaLighter metrics-and-logs stack, popular when a full SIEM is overkill
02 · Logs, SIEM & IOCs

SIEM Thinking: Correlation & Tuning

A correlation rule is a hypothesis: "if X and Y happen within Z time, something's wrong."

True PositiveThe alert was right, act on it
False PositiveThe alert was noise, tune it out
Alert fatigueToo many false positives and analysts stop trusting the queue, real threats start to slip through
TuningNarrowing a rule, thresholds, allow-lists, extra conditions, without losing real coverage
02 · Logs, SIEM & IOCs

What Is an IOC?

Indicator of Compromise, an observable artifact suggesting malicious activity happened.

File Hash

MD5 / SHA-1 / SHA-256

IP Address

A known-bad or C2 host

Domain / URL

Phishing or C2 infrastructure

Email address

A sender or reply-to used in a campaign

File name / path

Where malware drops or hides

Registry key

Persistence artifacts on Windows

02 · Logs, SIEM & IOCs

IOC vs. IOA

IOC: Indicator of Compromise

"What happened." A static artifact: a hash, a domain. Cheap for an attacker to change.

IOA: Indicator of Attack

"What's happening." A behavior: e.g. a process spawning a shell with an encoded command. Expensive for an attacker to change without changing their whole approach.

IOCs tell you an attacker was here. IOAs tell you one is here right now.
03

MITRE ATT&CK Introduction

A shared language for describing how adversaries actually operate

03 · MITRE ATT&CK

What Is MITRE ATT&CK?

Adversarial Tactics, Techniques & Common Knowledge, a free, globally-accessible knowledge base of real-world adversary behavior, maintained by MITRE.

  • Not a tool. A map, and a shared vocabulary the whole industry uses
  • Built from real intrusions, not theory
  • Used by vendors, threat intel teams, red teams, and blue teams alike
03 · MITRE ATT&CK

Tactics → Techniques → Procedures

Tactic
The WHYThe adversary's goal, e.g. Credential Access
Technique
The HOWThe method used, e.g. Brute Force (T1110)
Sub-technique
A specific variante.g. Password Spraying (T1110.003)
Procedure
The real-world instanceExactly how one threat actor or malware family did it
03 · MITRE ATT&CK

The Enterprise Matrix

Fifteen tactics. Each one answers: what is the adversary trying to achieve right now?

TA0043Reconnaissance
TA0042Resource Development
TA0001Initial Access
TA0002Execution
TA0003Persistence
TA0004Privilege Escalation
TA0005Stealthwas Defense Evasion
TA0112Defense Impairmentnew in v19
TA0006Credential Access
TA0007Discovery
TA0008Lateral Movement
TA0009Collection
TA0011Command & Control
TA0010Exfiltration
TA0040Impact

The framework is alive: in April 2026 (v19), MITRE split Defense Evasion into Stealth (TA0005, hiding) and Defense Impairment (TA0112, breaking your tools). Don't memorize this list, know how to look it up.

03 · MITRE ATT&CK

Case File 02: Same Incident, ATT&CK View

Recall the SSH log from Case File 01, five rapid failures against root, then a success.

Tactic
Credential AccessTA0006
Technique
Brute ForceT1110
Sub-technique
Password GuessingT1110.001

Repeated guesses against one account from one source = Password Guessing. Cycling one password across many accounts instead would be Password Spraying (T1110.003).

03 · MITRE ATT&CK

Why Blue Teams Use ATT&CK

  • Common language, a report, a rule, and a briefing can all point at the same technique ID
  • Coverage mapping, see which techniques you can actually detect, and where the gaps are
  • Threat-informed defense, prioritize based on what real adversaries in your sector do
  • Purple teaming, Red emulates a technique, Blue validates the detection, both sides close the loop

ATT&CK Navigator turns your coverage into a heat map, green where you can detect, red where you can't.

04

Hands-on Labs

Triage it, map it, then build the rule that would have caught it

04 · Hands-on Labs

Lab 01: Log Triage

A mixed batch of log lines, some routine, some not. Find the ones that matter, and say why.

Evidence set, mixed sources
108:02:11 sshd[1190]: Accepted password for maria from 10.0.4.21 port 55210 ssh2
208:41:37 nginx: 10.0.4.44 - "GET /dashboard HTTP/1.1" 200
309:14:02 sshd[2213]: Failed password for root from 185.221.104.7 port 51422 ssh2
409:14:05 sshd[2213]: Failed password for root from 185.221.104.7 port 51430 ssh2
509:14:09 sshd[2213]: Failed password for root from 185.221.104.7 port 51439 ssh2
609:14:47 sshd[2213]: Accepted password for root from 185.221.104.7 port 51502 ssh2
710:03:55 dns: query A api.internal-crm.local -> 10.0.2.15
810:47:12 powershell.exe -nop -w hidden -enc SQBFAFgAIAAoAE4AZQB3AC0ATw...
911:15:00 sshd[3390]: Accepted password for maria from 10.0.4.21 port 55880 ssh2

Task: flag the lines worth escalating. For each, say what makes it suspicious versus routine.

04 · Hands-on Labs

Lab 02: Attack Identification

Take what you flagged in Lab 01 and map each one to a MITRE ATT&CK tactic and technique.

ObservedTacticTechnique
Lines 3-6, SSH brute forceCredential AccessBrute Force · T1110.001
Line 8, encoded PowerShell??
Line 9, maria's second login??
Show suggested mapping

Line 8: Execution (TA0002), Command and Scripting Interpreter: PowerShell (T1059.001). The -enc flag and hidden window are also an attempt at Stealth (TA0005), obfuscating and hiding, not breaking tools, so this procedure reasonably maps to both.

Line 9: No tactic. Same user, same source IP, plausible hours, this is expected behavior. Correctly ruling something out is as important as flagging something.

04 · Hands-on Labs

Lab 03: Write a Detection Rule

Turn Case File 01 into a rule that would catch it automatically next time, using Sigma, a vendor-neutral rule format most SIEMs understand.

Sigma, base rule
title: SSH Failed Login
logsource:
product: linux
service: sshd
detection:
selection:
message|contains: 'Failed password'
condition: selection
level: low
Sigma, correlation rule
title: Repeated SSH failures from one source
correlation:
type: event_count
rules: [ssh_failed_login]
group-by: [src_ip]
timespan: 1m
condition: {gte: 5}

Task: adjust the threshold and timespan for your environment. What's a realistic false-positive scenario for this rule?

Show one reasonable answer

A shared jump box or NAT gateway where several real users legitimately fail-then-succeed could trip this rule.

Tune by: raising the threshold for known shared hosts, or adding a filter block that excludes trusted source IPs from the count.

Recap

Key Takeaways

  • Blue Team defends continuously, detect, respond, protect, recover, and govern the whole thing
  • Logs are the raw material; a SIEM turns them into something searchable and alertable
  • IOCs tell you something happened, IOAs tell you something is happening
  • MITRE ATT&CK gives everyone, vendors, analysts, reports, the same vocabulary for adversary behavior
  • Good detections start as a hypothesis, get tuned against real noise, and stay attached to a technique ID
Keep Going

Resources & Next Steps

attack.mitre.orgBrowse the matrix, techniques, and real procedure examples
sigmahq.ioThe Sigma rule spec and a public rule repository
Wazuh / Security OnionFree platforms to run your own mini-SOC at home
TryHackMe / LetsDefendGuided blue team exercises if you want more reps

Thanks, questions welcome.