2.3 Data Connectors & Ingestion

Key Takeaways

  • Connector families are service-to-service (API), agent-based via the Azure Monitor Agent (AMA), codeless (CCP), and custom; the legacy Log Analytics agent (MMA) is retired.
  • A Data Collection Rule (DCR) tells the AMA what to collect, how to transform it at ingestion time, and which workspace/table to target.
  • CEF via AMA lands in CommonSecurityLog, Syslog via AMA lands in Syslog, and DCR ingestion-time transformation can drop fields before billing.
  • The Advanced SIEM Information Model (ASIM) normalizes heterogeneous sources so one query/detection works vendor-agnostically across products.
  • The Content hub deploys Microsoft Sentinel solutions that bundle connectors with their analytics rules, workbooks, playbooks, and parsers.
Last updated: May 2026

Why This Matters for SC-200

Detections, hunting, and incident response are only as good as the data reaching Microsoft Sentinel and Microsoft Defender XDR. The 2026 skills outline explicitly lists ingesting Windows Security Events via AMA, Windows Event Forwarding (WEF), Syslog/CEF via AMA, Azure activities, threat indicators, and custom log tables. Expect "which connector / which agent" scenario questions.

Connector Types

Microsoft Sentinel data connectors fall into a few families:

  • Service-to-service (API) connectors — Native, low-effort connectors for Microsoft and major third-party clouds: Microsoft Entra ID, Microsoft Defender XDR, Office 365, Azure Activity, AWS, GCP. Typically enabled with a few clicks plus permissions.
  • Agent-based connectors — Use the Azure Monitor Agent (AMA) to collect logs from Windows and Linux machines (security events, Syslog, CEF, custom text/JSON logs).
  • Codeless / Codeless Connector Platform (CCP) connectors — Configuration-driven connectors for many SaaS/REST sources without custom code.
  • Custom connectors — Logs API / Logstash / Logic Apps / direct HTTP Data Collector for sources without a built-in connector.

The legacy Log Analytics agent (MMA/OMS) is retired. For SC-200 in 2026, the correct agent answer for collecting Windows or Linux logs is the Azure Monitor Agent (AMA), never the old Log Analytics agent.

Common Event Format, Syslog, and AMA / DCR

Many network and security appliances export logs over Syslog. When the appliance formats those Syslog messages using Common Event Format (CEF) — a structured key-value standard — Sentinel can parse them into the rich CommonSecurityLog table.

Modern collection works like this:

Appliance (CEF or Syslog)
   --> Linux forwarder VM running the Azure Monitor Agent (AMA)
        --> Data Collection Rule (DCR) defines what to collect & where
             --> Microsoft Sentinel / Log Analytics workspace
  • A Data Collection Rule (DCR) is the central configuration object that tells the Azure Monitor Agent what data to collect, how to filter/transform it, and which workspace/table to send it to.
  • CEF via AMA lands in CommonSecurityLog; generic Syslog via AMA lands in Syslog.
  • Windows Security Events via AMA collects domain/endpoint security events (and you choose All / Common / Minimal / Custom event sets in the DCR).
  • Windows Event Forwarding (WEF) funnels events from many Windows hosts to a collector server, which AMA then forwards — useful when you cannot install an agent on every endpoint.

Ingestion-time transformation in a DCR can drop or reshape data before billing — a common SC-200 cost-optimization and PII-filtering answer.

Loading diagram...
CEF/Syslog Ingestion via AMA and a Data Collection Rule

Custom Logs and Normalization (ASIM)

Custom logs

For applications without a built-in connector, ingest via custom text/JSON log collection through AMA + DCR, the Logs Ingestion API, Logstash, or Logic Apps. Custom data lands in a custom table (suffixed _CL). Choose the Auxiliary/Basic plan for high-volume, low-value custom logs to cut cost.

Advanced SIEM Information Model (ASIM)

Different sources name the same field differently (SourceIP, src_ip, ClientIP). The Advanced SIEM Information Model (ASIM) is Microsoft Sentinel's normalization layer. It provides:

  • Normalized schemas (for example, Authentication, Network Session, DNS, Process Event) with consistent field names.
  • Parsers that translate source-specific data into the normalized schema at query time.
  • Unifying/source-specific parsers so one analytics or hunting query (imAuthentication, _Im_NetworkSession, etc.) works across many products.

ASIM lets a single detection or hunting query run across heterogeneous sources without rewriting it per product — the SC-200 answer when a scenario asks for vendor-agnostic, source-independent detections.

Content Hub and Solutions

The Content hub is the in-product marketplace for Microsoft Sentinel solutions — packaged bundles that can include data connectors, analytics rules, hunting queries, workbooks, playbooks, parsers, and watchlists for a specific product or scenario (e.g., a vendor firewall solution). Installing the solution from the Content hub is the recommended, supported way to deploy a connector and its companion content together, instead of importing each artifact manually.

NeedCorrect choice
Collect Windows security events from serversWindows Security Events via AMA (DCR)
Ingest a firewall's CEF logsCEF via AMA to CommonSecurityLog
Drop noisy/PII fields before billingDCR ingestion-time transformation
One detection across many identity productsASIM normalized parsers
Deploy a vendor connector + its rules/workbooksInstall the solution from the Content hub
Test Your Knowledge

You must collect Windows Security event logs from on-premises domain controllers into Microsoft Sentinel in 2026, controlling exactly which event set is gathered. Which combination is correct?

A
B
C
D
Test Your Knowledge

A firewall sends high-volume Syslog messages formatted in Common Event Format. The SOC wants the structured data in CommonSecurityLog but must drop several verbose fields before they are billed. What should you configure?

A
B
C
D
Test Your Knowledge

An analytics rule must detect suspicious sign-ins consistently across Microsoft Entra ID, an on-premises VPN, and a third-party identity provider that each name fields differently. Which Microsoft Sentinel capability should the rule be built on?

A
B
C
D