Elastic Details Higher-Order Rules for Correlating Security Alerts

· Original article ↗

Summary

Elastic describes Higher-Order Rules that correlate endpoint, network, and observability alerts across entities and time to prioritize detections, with ES|QL examples, operational trade-offs, and reported alert-volume results.

Key points

  • Higher-Order Rules combine alerts with other alerts or telemetry, such as events and metrics, to surface patterns that may be more meaningful than isolated signals.
  • Elastic describes three design principles: correlate by shared entities, combine data sources, and account for time and prevalence.
  • Example ES|QL rules aggregate endpoint alerts by host or process tree; other examples correlate endpoint and network alerts using IP addresses to add process and user context.
  • First-seen rules use a lookback window to surface rare alerts, while observability correlations can pair resource spikes with security alerts.
  • Elastic reports that 65 detection rules generated nearly 8,000 alerts per day per production cluster; its Higher-Order detections produced about 30 alerts per day over a 15-day period, with tuning expected to reduce that to about 20.
  • The article cautions that correlation quality depends on base-rule quality and that scheduling delays and broad ES|QL queries can affect timeliness and cost; it recommends tuning base rules, scoping indices, and starting small.

Article Details

Defense Focus
Correlate alerts across entities, data sources, and time windows to improve detection confidence and prioritize triage.
Detection Methods
  • Correlate distinct endpoint alerts by host, process, or process tree; increase confidence when multiple event types, rule names, or file hashes converge.
  • Identify rare or newly observed alerts using first-seen times and lookback windows.
  • Join endpoint alerts with network security alerts using host and source IP fields, then enrich findings with process, command-line, parent-process, user, and device context.
  • Correlate Suricata network alerts with Elastic Defend network events using source port, source IP, and destination IP.
  • Correlate security alerts with sustained CPU or memory anomalies by host or process.
Data Sources
  • Endpoint alerts and process telemetry, including event codes, process trees, executable names, command lines, file hashes, host IDs, and user names
  • Network security alerts from Palo Alto Networks, FortiGate, and Suricata
  • Endpoint network events
  • System CPU and memory metrics
  • Email alerts and other integration alerts
Rule Types
  • ES|QL
  • EQL
  • Higher-Order correlation rules
  • Newly observed and first-seen rules
  • Hybrid correlation and first-seen rules
Platforms
  • Elastic Defend
  • Elastic Agent
  • Palo Alto Networks
  • FortiGate
  • Suricata
Defensive Actions
  • Tune base detections before enabling Higher-Order Rules to prevent noisy alerts from cascading into correlations.
  • Start with a small set of correlations; exclude inherently noisy tactics and deprioritize low-severity signals.
  • Scope index patterns to relevant datasets or use dataviews to reduce query costs.
  • Tune rule schedules and lookback windows to balance detection latency and performance.
  • Normalize alert severity and retire rules that disproportionately contribute to noisy outcomes.

Vendors

Products

Tools

Related Articles