The official AWS Certified Security - Specialty syllabus, organized by domain. Check off each point as you review it.
Inicia sesión para guardar tu progreso entre sesiones.
Design and implement monitoring and alerting solutions for an AWS account or organization
Monitoring requirements start from the workload type: GuardDuty analyzes CloudTrail, VPC flow logs and DNS logs with no extra setup, and protection plans extend it per resource (RDS Protection for Aurora and RDS logins, Lambda Protection, EKS Protection and Runtime Monitoring for containers and instances). Inspector, Macie and Security Hub CSPM cover vulnerabilities, sensitive data and controls.
GuardDuty foundational data sources ↗GuardDuty Runtime Monitoring ↗
Route 53 offers endpoint health checks (HTTP, HTTPS, TCP and string matching), calculated health checks and CloudWatch alarm-based health checks, useful for instances with only private IP addresses. Security groups must allow the health checkers through the managed prefix list, and alarms on health check status are created in US East (N. Virginia). Synthetics canaries mimic customers and detect unauthorized changes.
Types of Amazon Route 53 health checks ↗Configuring router and firewall rules for Amazon Route 53 health checks ↗
To centralize events, Security Hub CSPM replicates findings and control statuses to a home Region, GuardDuty uses the same delegated administrator account in every Region, an organization-based Config aggregator gives a read-only view without per-account authorization, and CloudWatch links source accounts to a monitoring account. EventBridge does not forward events received from another account to a third account.
Understanding cross-Region aggregation in Security Hub CSPM ↗Managing GuardDuty accounts with AWS Organizations ↗
Metric filters on CloudTrail logs in CloudWatch Logs, with alarms and SNS, alert on events such as root user usage; the CIS controls require the exact pattern. GuardDuty, Macie and Security Hub CSPM publish findings to EventBridge, and one rule in the administrator account covers the members. CloudWatch anomaly detection learns daily and weekly patterns instead of a static threshold.
Creating CloudWatch alarms for CloudTrail events: examples ↗Processing GuardDuty findings with Amazon EventBridge ↗
A conformance pack bundles AWS Config rules and remediation actions; deployed for the organization it reaches new accounts and members cannot modify it, but Config recording must be on. Config remediates with Automation documents, State Manager keeps nodes in a defined state on a schedule, and Security Hub CSPM central configuration enforces standards across accounts and Regions.
Managing Conformance Packs for AWS Config Across all Accounts in Your Organization ↗AWS Systems Manager State Manager ↗
Design and implement logging solutions
Each need has its source: CloudTrail data events (off by default) record Lambda Invoke and S3 object operations, but not requests with invalid credentials, which S3 server access logs do record. Resolver query logs keep DNS queries, Transit Gateway Flow Logs keep traffic through the transit gateway, and AWS WAF requires destinations whose names start with aws-waf-logs-.
An organization trail, created from the management or delegated administrator account, logs all accounts and member accounts cannot change it; the bucket policy aws:SourceArn uses the management account ID. The CloudWatch agent needs CloudWatchAgentServerPolicy and parameters prefixed AmazonCloudWatch-, API Gateway needs a cloudWatchRoleArn per Region, and CloudWatch Logs centralization copies new log data.
Creating a trail for an organization ↗Cross-account cross-Region log centralization ↗
Security Lake centralizes AWS, SaaS, on-premises and third-party data in S3 buckets in your account, normalized to OCSF in Apache Parquet. Native sources need no logging setup, except management events, which require a multi-Region organization trail. Custom sources must deliver OCSF in Parquet, and subscribers get data access (notifications) or query access (Athena) per source.
What is Amazon Security Lake? ↗Subscriber management in Security Lake ↗
CloudWatch Logs Insights analyzes logs in CloudWatch Logs with its query language, OpenSearch PPL or SQL (with JOINs across log groups) and speeds up searches with field indexes. Athena runs SQL over CloudTrail or flow logs in S3, and partition projection avoids adding partitions by hand. Security Hub CSPM insights group findings from several providers by an attribute.
Analyzing log data with CloudWatch Logs Insights ↗Create the table for CloudTrail logs in Athena using partition projection ↗
CloudWatch Logs transformers normalize events at ingestion with templates for AWS vended logs or Grok parsers. Security Lake normalizes data to OCSF in Parquet, OpenSearch Ingestion indexes it without servers, and OpenSearch Security Analytics applies Sigma rules with alerts. Managed Grafana combines several data sources in one panel.
Transform logs during ingestion - Amazon CloudWatch Logs ↗Security Analytics for Amazon OpenSearch Service ↗
VPC flow logs capture IP flows per interface, subnet or VPC, but not traffic to the Amazon DNS server; the pkt-srcaddr and pkt-dstaddr fields need a custom format, and the format cannot be changed after creation. Transit Gateway Flow Logs are created by the gateway owner, and Resolver query logging records queried names and can be shared with AWS RAM.
Flow log limitations - Amazon VPC ↗Resolver query logging - Amazon Route 53 ↗
Troubleshoot security monitoring, logging, and alerting solutions
Lambda writes logs only if its execution role has logs:CreateLogGroup, CreateLogStream and PutLogEvents. API Gateway needs a role in cloudWatchRoleArn in each Region. Route 53 health checks require allowing the route53-healthchecks prefix list, and CloudFront legacy logging needs ACLs enabled and a customer managed key when SSE-KMS is used.
Set up CloudWatch logging for REST APIs in API Gateway ↗Configure standard logging (legacy) - Amazon CloudFront ↗
When CloudWatch agent logs are missing: an i/o timeout in private subnets calls for interface endpoints with private DNS; after editing the configuration you must run fetch-config; without log_group_name the group is named after file_path. Flow logs need a role that trusts vpc-flow-logs.amazonaws.com, and CloudTrail needs permissions in the bucket policy and the KMS key policy.
Troubleshooting the CloudWatch agent - Amazon CloudWatch ↗Amazon S3 bucket policy for CloudTrail - AWS CloudTrail ↗