SMART WORLDJOURNALSubmit Manuscript
← All technology news

Digital Infrastructure

Amazon GuardDuty Enhances Threat Detection Customization to Combat Alert Fatigue

Amazon Web Services updates GuardDuty with Custom Detection Rules, expanded filtering, and AI-powered investigations to help security teams reduce noise, tailor alerts, and streamline investigations across multi-account environments.

NEWS

Security teams managing cloud environments often face alert fatigue due to excessive findings that are not contextualized to their specific operations. Amazon GuardDuty, AWS's threat detection service, continuously monitors AWS data sources like CloudTrail logs, VPC Flow Logs, and DNS query logs to identify suspicious activity and generate findings. While this managed detection framework maintains consistent baseline monitoring, recent features now empower users to customize detection criteria to better fit their environments, significantly reducing false positives and unnecessary alerts.

The centerpiece of this customization is the introduction of Custom Detection Rules, a curated library of over 90 rules that monitor for behaviors specific to AWS CloudTrail management events and GuardDuty Runtime Monitoring events. These rules, mapped to recognized threat tactics cataloged by the AWS Customer Incident Response Team and MITRE ATT&CK® techniques, allow teams to opt-in to signals relevant to their environment. Users may first run these rules in dry run mode for a 14-day evaluation period to observe potential noise through CloudWatch metrics before activating them to generate real findings. This approach ensures that rules can be scoped and refined, minimizing irrelevant alerts and optimizing detection accuracy.

Additionally, GuardDuty now supports sophisticated threat entity lists that can include IP addresses, CIDR ranges, domain names, and even SHA-256 file hashes. Integration of these lists extends coverage across network and runtime telemetry, improving the detection of known malicious activity. Custom threat intelligence can be uploaded and maintained in Amazon S3, allowing organizations to align GuardDuty findings with internal or external intelligence feeds.

To further reduce noise, advanced suppression rules have been enhanced to support wildcards and over a thousand new filter fields derived from findings. These granular filters enable precise archiving of expected benign findings, such as those generated by routine vulnerability scans or trusted service accounts, preventing these from cluttering analyst workflows or propagating to SIEM tools like AWS Security Hub. Crucially, suppression archives findings rather than deleting them, preserving auditability and allowing reversal if needed.

For investigations, AWS has rolled out AI-powered GuardDuty Investigation, which synthesizes related activity over a 90-day window alongside contextual threat data to provide risk scores, confidence levels, detailed summaries, and mapped MITRE ATT&CK classifications. This automated analysis can be accessed via console, CLI, or API to prioritize incidents and suggest remediation steps, saving analysts considerable time previously spent manual data gathering and correlation.

These features are available to customers with GuardDuty enabled in supported regions, including the AWS commercial and GovCloud (US) regions, though some capabilities like AI-powered investigation and entity lists have regional limitations. Organizations managing multiple AWS accounts are advised to use the delegated administrator model for centralized configuration and rule management, enabling consistent enforcement across their environment.

By combining tailored detection, precise filtering, and AI-accelerated analysis, Amazon GuardDuty aims to alleviate alert fatigue, enhancing security posture and operational efficiency for AWS customers. However, effective use requires initial investment in tuning and ongoing review to adapt to changing environments and threat landscapes.

While these capabilities improve detection relevance and analyst productivity, they do not eliminate the need for expert security teams to interpret findings and respond appropriately. Additionally, data residency considerations arise with AI-powered investigations using cross-region inference, which may process analysis outside the originating region.

Source & further reading

Reduce alert fatigue by customizing threat detection features in Amazon GuardDuty ↗

Source published 2026-11-10. Claims and limitations should be checked against the original announcement or research.