A security operations team in Microsoft Sentinel must alert on a specific privileged sign-in pattern with the lowest possible detection latency, evaluating events almost as soon as they arrive in the workspace rather than waiting for the next scheduled run. The detection logic queries a single table and needs no enrichment from other tables. Which analytics rule type meets this requirement?
- AA scheduled analytics rule configured with the shortest supported query interval and lookback period
- BA Microsoft security analytics rule that creates incidents from connected Defender product alerts
- CA near-real-time (NRT) analytics rule that runs once per minute against incoming events Correct
- DA Fusion rule that correlates multistage activity into a single high-fidelity incident
Why A is wrong: A scheduled rule runs on a fixed timer, so even the shortest interval still introduces minutes of delay and cannot match the near-real-time evaluation the requirement demands.
Why B is wrong: A Microsoft security rule only forwards alerts already raised by Defender connectors into incidents; it does not run custom Kusto Query Language detection logic against a table.
Why C is correct: An NRT rule evaluates events roughly every minute as they ingest, giving the lowest detection latency for a single-table query that needs no cross-table enrichment.
Why D is wrong: Fusion uses machine learning to stitch multistage attacks across many signals; it is not a low-latency single-pattern detector and its logic cannot be authored by the team.