Skip to content
Tutorials Center
My learning

Topics

Lesson 3 of 37 ยท Getting Started

The SIEM Landscape: Splunk, Elastic, Sentinel and Wazuh

Compare popular SIEM platforms, their query languages and licensing models, and learn how to choose one. Read-only SPL, KQL and ES|QL examples plus a Python stand-in.

There is no single "best" SIEM. Products differ in pricing, query language, hosting and how much work they leave to you. The concepts in this course carry across all of them, so learn the ideas first and the syntax second.

None of these products run on this site. The query examples below are read-only and labelled with the product they belong to; the runnable code uses a Python stand-in that shows the same logic.

Common platforms

Platform Type Query language Notes
Splunk Enterprise Security Commercial, self-hosted or cloud SPL Very flexible search, priced by data volume or workload
Elastic Security Open core, self-hosted or cloud KQL, EQL, ES|QL Built on Elasticsearch, strong detection rules library
Microsoft Sentinel Cloud-native SaaS KQL Tight Azure and Microsoft 365 integration, priced per GB
Google SecOps (Chronicle) Cloud-native SaaS YARA-L Large retention, normalized to the UDM model
IBM QRadar Commercial AQL Offense-based correlation
Wazuh Open source Dashboards + rules XML Agents, file integrity, vulnerability detection
Graylog, Security Onion Open source Own syntax Popular for small teams and labs

The same question in three languages

Question: show failed logins per source IP in the last hour, only where there are at least five.

Splunk (SPL), read-only:

TEXT
index=auth action=failure earliest=-1h
| stats count by src_ip
| where count >= 5
| sort - count

Microsoft Sentinel (KQL), read-only:

TEXT
SigninLogs
| where TimeGenerated > ago(1h) and ResultType != 0
| summarize failures = count() by IPAddress
| where failures >= 5
| order by failures desc

Elastic (ES|QL), read-only:

TEXT
FROM logs-auth-*
| WHERE event.outcome == "failure" AND @timestamp > NOW() - 1 hour
| STATS failures = COUNT(*) BY source.ip
| WHERE failures >= 5
| SORT failures DESC

All three do the same four steps: filter, group, threshold, sort. Once you can think in "pipelines of steps", switching products takes days, not months.

A pipeline stand-in you can run

Python Filter, group, threshold, sort in Python

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Deployment models

  • Self-hosted: you run servers and storage. Full control, you own patching and scaling.
  • Cloud (SaaS): the vendor runs it; faster to start, data leaves your network, and costs scale with ingest.
  • Hybrid: sensitive logs stay local, others go to cloud.
  • MDR / managed SIEM: a provider operates the SIEM and the 24/7 watch for you.

Pricing models decide behavior

How a vendor charges changes what you collect:

Model Meaning Side effect
Per GB ingested per day Pay for volume Teams drop noisy logs to save money
Per workload / compute Pay for searches and nodes Heavy searching costs more
Per endpoint / host Flat price per agent Cheap for small estates
Open source No license, you pay in people Needs engineering skill
Python Estimate monthly log cost

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The output shows a common truth: a few sources produce most of the volume. Decide per source whether the security value justifies the cost; do not just drop data blindly.

How to choose

  1. List your must-have sources (cloud, endpoints, identity) and check connectors exist.
  2. Check who will write rules and what language they already know.
  3. Estimate ingest per day and test the pricing at that size.
  4. Decide self-hosted versus cloud based on data residency rules.
  5. Run a proof of concept on real log samples before buying.

Learn one platform deeply plus the shared concepts: Sigma rules (chapter 4) are a vendor-neutral format that converts to many of these languages.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Which query language belongs to Microsoft Sentinel?

    Sentinel uses KQL (Kusto Query Language). SPL is Splunk, AQL is QRadar and YARA-L is Google SecOps.
  2. 2. What is the effect of per-GB pricing on teams?

    Volume pricing pushes teams to decide which sources are worth the money, sometimes cutting useful data.
  3. 3. What do SPL, KQL and ES|QL queries have in common?

    Different syntax, same logical steps, so concepts transfer between products.
  4. 4. What is a Managed Detection and Response (MDR) service?

    MDR outsources detection and response work, often including a hosted SIEM.

Summary

  • Splunk, Elastic, Sentinel, Chronicle, QRadar and Wazuh share the same architecture with different languages and pricing.
  • Queries are pipelines: filter, group, threshold, sort.
  • Pricing shapes what you collect, so measure volume per source.
  • Choose using connectors, skills, cost at your volume and data residency; test with a proof of concept.

Next: Build a Safe Practice Lab