Skip to content
Getting Digital

SIEM (Security Information and Event Management)

Also: SIEM, log management, security analytics, event correlation

A SIEM collects logs and events from across an organisation's systems into one searchable store, correlates them into alerts, and retains them for investigation and compliance.

Our take. A SIEM is priced by how much you feed it and valued by how little of that you need, and most deployments get the ratio backwards. Ingest the identity provider, the endpoints, the firewall and the cloud control plane, write detections against those, and add sources only when a detection needs them. Everything else is a storage bill with a security label.

Every system writes logs, and an attack leaves traces across several of them: a login from a new country in the identity provider, a new process on an endpoint, an outbound connection in the firewall, a permission change in the cloud console. No single log shows the attack. A SIEM puts them in one place, normalises the fields so a username means the same thing in each, and runs rules that fire when a combination appears: the new-country login followed within minutes by a permission change is a detection; either alone is noise. Retention lets an analyst ask, months later, what else that account did, which is the question every investigation eventually reaches.

  • Identity provider logs: who signed in, from where, with what factor. The single most valuable source and the cheapest.
  • Endpoint telemetry: processes, network connections and file changes on each device. High volume, high value.
  • Firewall and proxy logs: who talked to whom, especially outbound. Where command-and-control traffic shows.
  • Cloud control plane: every configuration change in the cloud account, which is where modern attackers do their damage.
  • Application and database logs: valuable for specific detections and expensive in bulk; add per detection, not wholesale.

The cost model shapes the security model

Most SIEMs charge by volume ingested or by events per second, so every source is a budget decision and teams routinely leave out the expensive ones, which are often the endpoints and the applications. The consequence is a system that knows a login happened and not what the user did afterwards. Decide what you need to detect, then ingest what those detections require, then argue about the bill. Ingesting everything and writing detections later produces the bill without the detections.

The analyst certifications live here. Microsoft's operations exam is built around its own SIEM and the query language that interrogates it; the vendor-neutral analyst paper tests reading correlated events and deciding which alert matters. Both assume the same skill: taking a query, saying what it would return, and noticing what it would miss. That is also the skill that separates a SIEM that finds things from one that stores them.

In practice

A company ingests eight hundred sources into its SIEM and has a detection rule set copied from a vendor template. An attacker phishes a credential, logs in from a new country, adds a mail forwarding rule, and reads the finance team's inbox for three weeks. The identity log recorded the login, the mail system recorded the forwarding rule, and both were in the SIEM. No rule joined them. The rewrite starts from the attack rather than the data: a detection for a new-country sign-in followed by a mailbox rule change within an hour, another for a sign-in without the expected second factor, another for a permission grant by an account that has never granted one. Six detections, written against three sources, catch the next attempt in eleven minutes.

Often confused with

Intrusion Detection and Prevention (IDS/IPS)
An intrusion detection system produces alerts from network or host activity. A SIEM is where those alerts land alongside every other log, and where they gain the context that turns an alert into a case.
Security Operations Centre (SOC)
The SIEM is the tool; the security operations centre is the team that writes its detections and works its alerts. A SIEM without a team is an archive.
Incident Response
A SIEM tells you something may have happened and lets you investigate what. Incident response is what the organisation does about it once it is confirmed, and it starts where the SIEM's job ends.

Key takeaways

  • An attack leaves traces in several logs; the SIEM is where they meet and become one detection.
  • Identity, endpoint, firewall and cloud control plane first; everything else per detection.
  • Decide what to detect, then ingest what that requires; not the other way round.

Certifications that test this

Vendor exams whose syllabus covers this concept — facts, cost and a preparation path on each page.

More courses from these shelves

A rotating selection from the course directory, drawn from the subcategories where this concept is taught rather than picked for it. Details, price and the provider link are on the course page.

ISO/IEC 42001: Artificial Intelligence Management System

ISO/IEC 42001: Artificial Intelligence Management System is a comprehensive course designed for professionals looking t…

Udemy

C_THR82: SuccessFactors Performance & Goals Implementation

Are you ready to pass the SAP Certified Associate - SAP SuccessFactors Performance and Goals (C_THR82) exam and take yo…

Udemy

Networking Full Course & Network + certification

This Class of Full Networking Fundamentals, will be fully illustrated with video lessons and sample to which it will ma…

Udemy

Python And Django Framework For Beginners Complete Course

Learn Python From Scratch Beginner to Expert Python.Start from the Python basics and go all the way to creating your ow…

Udemy

The Ultimate AWS Networking Training Course: All In One

Unlock the Future of Cloud: Master AWS Networking and Propel Your Career Forward!In a world powered by the cloud, Amazo…

Udemy

Salesforce Certified Data Cloud Consultant Practice Exams

Get certified with Salesforce Data Cloud Consultant certification by practicing actual exam type questions. This Course…

Udemy

FAQ

Do I need a SIEM as a small organisation?
Probably not a heavy one. Central logging of the identity provider, the endpoints and the cloud account into something searchable, with a handful of detections, covers most of the value, and several vendors bundle that with their endpoint product. A full SIEM makes sense when there is someone whose job is to work its alerts.
SIEM or plain log management?
Log management stores and searches logs. A SIEM adds normalisation across sources, correlation rules that join events into alerts, and case handling. Many products do both, and many organisations pay for the second and use only the first.
How long should logs be kept?
Long enough to investigate an attack discovered late, which in practice means months for the searchable tier and longer in cheap cold storage for the sources regulation names. The exact figure is set by the slowest-discovered attack you can imagine and by your compliance obligations, not by the vendor's default.

Sources

The primary text this definition rests on. Read it before you trust ours.

Last reviewed 13 September 2026 · Getting Digital