Skip to content
StrikeCyberStrikeCyber
Field Manual · Concept

SIEM

Reviewed November 14, 2026Concept 2 min read

Quick answer

A SIEM (security information and event management system) collects logs from across an organization, such as servers, firewalls, identity systems and cloud services, and raises alerts when it sees signs of an attack. It is only as good as the logs it receives and the detection rules written for it, which is why SIEMs often miss real intrusions.

On this page

Almost every organization that suffers a serious breach had logs that showed the attacker. The problem is rarely that nothing was recorded; it is that nobody was alerted, or the alert was lost among thousands of others. A SIEM is the system meant to close that gap.

01

How It Works

A SIEM does four things: it collects logs from many sources, normalizes them into a common format, correlates events across sources using detection rules, and raises alerts for analysts to investigate. Modern SIEMs add threat intelligence, user behavior analytics and automated response (SOAR).

Log sourceWhat it reveals
Identity (Active Directory, Entra ID)Logins, password sprays, privilege changes, suspicious Kerberos activity
Endpoints and EDRMalware, credential theft, suspicious processes
Firewalls, proxies and DNSCommand and control traffic, data exfiltration
Cloud and SaaS (Microsoft 365, AWS, Azure)Risky sign-ins, inbox rules, mass downloads, configuration changes
Servers and applicationsWeb attacks, database access, service creation
02

Why It Matters

A SIEM is a large investment that often underdelivers. Common gaps: key sources are not collected (domain controller logs and cloud audit logs are frequently missing), detection rules are vendor defaults that never fired in testing, alerts are too noisy to act on, and nobody has measured what an attacker would actually trigger. Attackers who use legitimate tools and valid accounts can move through a network without tripping any default rule.

03

How to Test for It

  1. List the log sources the SIEM actually receives and compare them with your critical systems.
  2. Map existing detection rules to MITRE ATT&CK techniques to see what is covered on paper.
  3. Run realistic techniques safely (password spraying, Kerberoasting, lateral movement, data staging) and record what was logged and what alerted.
  4. Measure time to detect and time to respond, and whether analysts escalated correctly.
  5. Tune or write rules for the gaps, then retest.
04

How to Reduce the Risk

Prioritize the logs attackers most often leave in: identity, endpoint, email and cloud audit logs. Write detections for the techniques used in real intrusions rather than relying on defaults, retain logs long enough to investigate (at least 90 days searchable, longer archived), and test detection regularly with purple teaming so improvements are measured rather than assumed.

05

Regulatory and Framework Context

For US organizations, centralized logging and monitoring are required by frameworks such as NIST SP 800-53 (the AU control family), PCI DSS Requirement 10 and SOC 2, and are central to the NIST Cybersecurity Framework's Detect function. Our SOC assessment and detection testing shows what your SIEM would catch.

Latest intel on this topic

Updated hourly from the daily briefing.

Common questions

What is the difference between a SIEM and EDR?

EDR watches individual devices in depth and can stop malicious activity on them. A SIEM collects data from everywhere, including EDR, firewalls, identity and cloud, and correlates it to find attacks that no single source shows. Most organizations need both.

What are common SIEM products?

Widely used SIEMs include Microsoft Sentinel, Splunk Enterprise Security, Google Security Operations, Elastic Security, IBM QRadar and Sumo Logic. Many organizations use one through a managed security service provider (MSSP).

Do we need a SIEM?

Organizations with more than a handful of servers or any compliance obligation usually need centralized logging and alerting. Smaller organizations often get better value from EDR plus a managed detection service that includes log monitoring.

How do we know our SIEM works?

Test it: run realistic attack techniques in a controlled way and check what was logged, what alerted, and how fast someone responded. This is what purple teaming and adversary simulation measure.

Change log

  1. November 14, 2026First published.
(877) 657-8496