AWS Trusted Advisor: The Complete Guide to Cost, Security, and Performance Checks
DevOps / SRE Engineer & CEO, Cloud DevOps Solutions
AWS Trusted Advisor is a built-in service that inspects your AWS account and flags issues across five categories: cost optimisation, security, fault tolerance, performance, and service limits. Think of it as an automated auditor that continuously checks whether you are following AWS best practices.
Most teams know Trusted Advisor exists but treat it as a dashboard they glance at once a quarter. That is a mistake. When used properly, with the right tier, the right automation, and the right integration into your review process, Trusted Advisor catches problems weeks before they become incidents or surprise bills.
This guide explains what Trusted Advisor checks, how the tiers differ, how to automate recommendations, and how it fits into a broader cloud governance strategy.
What AWS Trusted Advisor does
Trusted Advisor runs a set of predefined checks against your AWS account. Each check evaluates a specific best practice and returns one of four statuses:
- Green (OK): the resource passes the check
- Yellow (warning): investigation recommended
- Red (action recommended): the resource fails the check and you should fix it
- Grey (excluded): the check was suppressed or is not applicable
Checks are grouped into five categories. Some are available on every AWS account; others require a Business, Enterprise On-Ramp, or Enterprise support plan.
The five check categories
1. Cost optimisation
Cost checks look for resources you are paying for but not using, or resources that could be cheaper with a different purchasing model. Key checks include:
- Low-utilisation EC2 instances: flags instances where average CPU utilisation has been below 10% for 14 days. A single m5.xlarge running at 3% costs roughly $140/month for nothing.
- Idle load balancers: Application Load Balancers with no healthy targets or no incoming requests.
- Unassociated Elastic IP addresses: each unused EIP costs $3.65/month. Across 50 accounts, that adds up.
- Underutilised EBS volumes: volumes with low read/write operations, especially provisioned IOPS volumes you are paying a premium for.
- Amazon RDS idle instances: database instances with no connections for an extended period.
- Savings Plans and Reserved Instance coverage: highlights On-Demand spend that could shift to a commitment-based pricing model for 30-60% savings.
- S3 incomplete multipart uploads: orphaned upload parts that accumulate storage charges quietly.
Cost checks alone have saved teams we work with anywhere from 15% to 40% on their monthly bill. The recommendations are specific: they name the resource, the region, and the potential saving. See our cloud cost optimisation service for how we apply these findings systematically.
2. Security
Security checks identify configuration gaps that expose your account to risk:
- S3 bucket permissions: flags buckets with public access or overly permissive ACLs. This check has prevented more data breaches than any other single control.
- Security groups with unrestricted access: rules allowing 0.0.0.0/0 on sensitive ports (22, 3389, 3306, 5432).
- IAM use: whether the root account has MFA enabled, whether IAM users have MFA, and whether access keys are rotated.
- CloudTrail logging: whether CloudTrail is enabled in all regions. Without it, you have no audit trail.
- EBS and RDS encryption: flags unencrypted volumes and database instances.
- AWS Config: whether AWS Config is recording resource changes.
- Exposed access keys: checks public code repositories for AWS access keys associated with your account.
For teams with compliance requirements (SOC 2, ISO 27001, HIPAA), these checks provide a baseline. They are not a replacement for a full security assessment, but they catch the most common misconfigurations that auditors flag. Our cloud security consulting builds on these foundations.
3. Fault tolerance
Fault tolerance checks evaluate whether your architecture can survive failures:
- Amazon RDS multi-AZ: flags single-AZ database instances that would cause downtime during an AZ failure.
- Amazon RDS backups: instances with automated backups disabled.
- EC2 Availability Zone balance: whether your instances are distributed across multiple AZs.
- ELB cross-zone load balancing: whether load is distributed evenly across AZs.
- Auto Scaling group health check: whether health checks are configured to replace unhealthy instances.
- VPN tunnel redundancy: flags VPN connections with only one active tunnel.
- Route 53 health checks and failover: whether DNS failover is configured for critical endpoints.
These checks matter most for production workloads. A single-AZ RDS instance is fine for a development database but a ticking time bomb in production.
4. Performance
Performance checks identify bottlenecks and configuration issues that degrade application speed:
- High-utilisation EC2 instances: instances running at sustained high CPU, suggesting they need to be right-sized upward or the workload distributed.
- CloudFront configuration: whether content delivery is optimised (header forwarding, TTL settings, alternate domain names with HTTPS).
- EC2 to EBS throughput optimisation: whether the instance type’s EBS throughput matches the volume’s capability.
- Large number of rules in a security group: security groups with excessive rules slow down network processing.
5. Service limits (quotas)
Service limit checks warn you before you hit AWS account quotas:
- VPCs per region
- Elastic IPs per region
- IAM policies, roles, and groups
- Auto Scaling groups and launch configurations
- RDS instances and storage
- EBS volumes and snapshots
Hitting a service limit during a scaling event or deployment is one of the most frustrating failures in AWS. These checks give you lead time to request increases. Trusted Advisor flags usage above 80% of the limit as a warning and above 100% as an error.
Trusted Advisor tiers: Basic vs Business vs Enterprise
Not all checks are available on every support plan. This is the most common source of confusion.
| Feature | Basic / Developer | Business | Enterprise On-Ramp | Enterprise |
|---|---|---|---|---|
| Core security checks | 6 checks | All | All | All |
| Service limit checks | All | All | All | All |
| Full check library | No | Yes (all 300+) | Yes | Yes |
| API access | No | Yes | Yes | Yes |
| CloudWatch integration | No | Yes | Yes | Yes |
| Programmatic refresh | No | Yes | Yes | Yes |
| Trusted Advisor Priority | No | No | Yes | Yes |
| Weekly email notifications | No | Yes | Yes | Yes |
Basic and Developer plans
You get six core security checks and all service limit checks. The security checks cover S3 bucket permissions, security groups, IAM use, MFA on root, CloudTrail, and EBS public snapshots. These are valuable but limited.
If you are on a Developer plan and want the full check library, upgrading to Business support is the only path. Business support starts at the greater of $100/month or 10% of your monthly AWS charges for the first $10,000.
Business plan
Unlocks the full library of 300+ checks across all five categories. You also get API access, which is essential for automation (covered below), and the ability to programmatically refresh check results instead of waiting for the weekly cycle.
Enterprise On-Ramp and Enterprise
Adds Trusted Advisor Priority, a curated view where a Technical Account Manager (TAM) helps prioritise which recommendations to act on first. This matters for large organisations with thousands of findings where “fix everything” is not a realistic approach.
Trusted Advisor Priority
Trusted Advisor Priority, available on Enterprise On-Ramp and Enterprise support plans, aggregates recommendations from Trusted Advisor, AWS Health, and AWS Well-Architected Tool into a single prioritised list. Your TAM reviews and ranks the findings based on your account’s specific context.
Priority recommendations include:
- Risk level: critical, warning, or informational
- Affected resources: specific ARNs
- Recommended actions: step-by-step remediation
- Acknowledgement workflow: mark items as in progress, resolved, or accepted risk
For organisations managing dozens or hundreds of AWS accounts through AWS Organizations, Priority rolls up findings across all member accounts, giving the platform team a single view.
How to access Trusted Advisor
AWS Console
Navigate to the Trusted Advisor console (Support > Trusted Advisor in the navigation). The dashboard shows a summary across all categories with green/yellow/red counts.
AWS CLI
# List all available checks
aws support describe-trusted-advisor-checks --language en
# Get results for a specific check
aws support describe-trusted-advisor-check-result \
--check-id "Qch7DwouX1" \
--region us-east-1
# Refresh a check (Business+ plans only)
aws support refresh-trusted-advisor-check \
--check-id "Qch7DwouX1"
Note: the Trusted Advisor API is available only through the us-east-1 endpoint, regardless of which region your resources are in.
AWS SDKs
All AWS SDKs support the Support service client. Use describeTrustedAdvisorChecks to list checks and describeTrustedAdvisorCheckResult to get findings.
Automating Trusted Advisor with EventBridge and Lambda
Checking Trusted Advisor manually defeats the purpose. Here is how to automate it.
Step 1: Enable Trusted Advisor as an EventBridge source
Trusted Advisor publishes events to Amazon EventBridge when check statuses change. The event detail type is Trusted Advisor Check Item Refresh Notification.
Step 2: Create an EventBridge rule
{
"source": ["aws.trustedadvisor"],
"detail-type": ["Trusted Advisor Check Item Refresh Notification"],
"detail": {
"status": ["WARN", "ERROR"]
}
}
This rule triggers only on warnings and errors, not on passing checks.
Step 3: Route to a Lambda function
The Lambda function receives the event, extracts the check name, status, affected resource, and recommended action, then routes it:
- Slack/Teams notification for warnings
- PagerDuty/Opsgenie alert for errors on critical checks (security group open to the world, MFA disabled on root)
- Jira/Linear ticket for cost recommendations above a threshold
Step 4: Auto-remediation (optional)
For certain checks, you can automate the fix:
- Unassociated Elastic IPs: release them automatically
- S3 public access: apply
PublicAccessBlockconfiguration - Security groups with 0.0.0.0/0 on SSH: remove the offending rule and notify the team
Auto-remediation should be limited to low-risk actions with clear rollback paths. Never auto-remediate without logging what changed.
Using AWS Config rules as a complement
Trusted Advisor runs checks periodically; AWS Config evaluates rules continuously against configuration changes. For real-time enforcement, pair Trusted Advisor with AWS Config managed rules:
| Trusted Advisor check | AWS Config managed rule |
|---|---|
| S3 public access | s3-bucket-public-read-prohibited |
| Security group unrestricted | restricted-ssh, restricted-common-ports |
| EBS encryption | encrypted-volumes |
| RDS encryption | rds-storage-encrypted |
| CloudTrail enabled | cloud-trail-cloud-watch-logs-enabled |
| Root MFA | root-account-mfa-enabled |
AWS Config catches issues within minutes of a configuration change. Trusted Advisor provides the broader account-level view and the cost/performance checks that Config does not cover.
Trusted Advisor and the AWS Well-Architected Framework
The AWS Well-Architected Framework defines six pillars: operational excellence, security, reliability, performance efficiency, cost optimisation, and sustainability. Trusted Advisor maps directly to four of them:
| Well-Architected pillar | Trusted Advisor category |
|---|---|
| Security | Security checks |
| Reliability | Fault tolerance checks |
| Performance efficiency | Performance checks |
| Cost optimisation | Cost optimisation checks |
When we run an AWS Well-Architected Review for a client, Trusted Advisor findings are the starting point, not the finish line. The review goes deeper: architecture patterns, operational procedures, deployment strategies, and organisational practices that Trusted Advisor cannot evaluate.
However, if Trusted Advisor shows red across multiple categories, it is a strong signal that the account needs a structured review rather than spot fixes.
Trusted Advisor across multiple accounts with AWS Organizations
For organisations using AWS Organizations, Trusted Advisor supports an organisational view that aggregates findings across all member accounts. This requires:
- The management account or a delegated administrator account
- Business or Enterprise support on each member account (checks run per-account, not per-organisation)
- Organisational view enabled in Trusted Advisor settings
The organisational view generates reports in CSV or JSON format, stored in an S3 bucket you designate. Platform teams use these reports to:
- Identify which accounts have the most critical findings
- Track remediation progress over time
- Enforce baseline compliance standards across all accounts
- Produce executive dashboards for leadership
Common Trusted Advisor findings and how to fix them
Idle resources (cost)
Finding: “X EC2 instances have low utilisation”
Fix process:
- Verify the instances are genuinely idle (check CloudWatch metrics for CPU, memory via CloudWatch Agent, and network)
- For dev/test instances: stop them outside business hours with Instance Scheduler or an EventBridge rule
- For production: right-size to a smaller instance type or migrate the workload to a container (Fargate or EKS)
- For batch workloads: move to Spot instances
See our guide on EKS cost optimisation for container-specific strategies.
Unrestricted security groups (security)
Finding: “Security group X allows unrestricted access on port 22”
Fix process:
- Identify who needs SSH access (developers, CI runners, bastion host)
- Replace 0.0.0.0/0 with specific CIDR ranges or security group references
- Better: remove SSH access entirely and use AWS Systems Manager Session Manager for shell access. No open ports, full audit trail, IAM-based access control.
Single-AZ RDS instances (fault tolerance)
Finding: “RDS instance X is not configured for multi-AZ”
Fix process:
- For production databases: enable multi-AZ. The failover adds approximately 2-4 minutes of downtime during an AZ failure instead of hours of manual recovery.
- For development: accept the risk, but document it. Mark the Trusted Advisor finding as “accepted risk” so it does not clutter the dashboard.
- For critical databases: consider Aurora with read replicas across three AZs for sub-minute failover.
Service limit warnings (quotas)
Finding: “You are using 85% of your VPC limit in eu-west-1”
Fix process:
- Request a quota increase through Service Quotas before you hit the wall
- Automate quota monitoring with CloudWatch alarms on
ServiceQuotametrics - Review whether you actually need that many VPCs, or whether a shared VPC model with Transit Gateway would reduce the count
Trusted Advisor vs third-party tools
Trusted Advisor is free (with your support plan) and tightly integrated with AWS. But it has limits:
| Capability | Trusted Advisor | Third-party tools (e.g. Datadog, Vantage, CloudHealth) |
|---|---|---|
| AWS-native, no setup | Yes | No (requires agent or API integration) |
| Cost recommendations | Basic (idle, low-utilisation) | Advanced (container right-sizing, cross-service correlation) |
| Security checks | Surface-level best practices | Deep analysis (CIS benchmarks, custom policies) |
| Multi-cloud | AWS only | AWS + Azure + GCP |
| Custom checks | No | Yes |
| Historical trending | Limited | Full |
| Kubernetes-aware | No | Yes (some tools) |
For teams that are AWS-only and want a no-cost starting point, Trusted Advisor is sufficient for the basics. For organisations with complex multi-account environments, container workloads, or multi-cloud infrastructure, a third-party tool complements Trusted Advisor rather than replacing it.
Best practices for getting value from Trusted Advisor
Review findings weekly, not quarterly
Schedule a 30-minute weekly review of Trusted Advisor findings. Focus on new red items. Assign each one to a team member with a deadline.
Automate notifications
Do not rely on the console. Set up EventBridge rules to push warnings and errors to your team’s communication channel. The weekly email notification from AWS is a good backup but not a primary workflow.
Suppress findings you have accepted
If you have intentionally configured something that Trusted Advisor flags (a public S3 bucket for a static website, a single-AZ dev database), suppress the finding. A dashboard full of yellow items that you have already reviewed trains people to ignore it.
Track savings
When you act on a cost recommendation, record the before and after spend. This builds the business case for continued investment in cloud governance and demonstrates ROI to leadership.
Combine with AWS Well-Architected Tool
Run the Well-Architected Tool assessment alongside Trusted Advisor findings. The Well-Architected Tool asks questions about your architecture and practices; Trusted Advisor provides the evidence. Together, they give a complete picture.
Use the API for dashboards
For organisations with multiple accounts, build a dashboard that pulls Trusted Advisor data via the API, aggregates it, and shows trends over time. Open-source tools like Prowler complement this with deeper security analysis.
When Trusted Advisor is not enough
Trusted Advisor checks for known patterns. It does not:
- Analyse your application architecture for anti-patterns
- Evaluate your CI/CD pipeline for deployment risks
- Review your Kubernetes resource requests and limits
- Assess your team’s operational maturity
- Benchmark your spend against industry peers
These require a structured review by someone who understands both the technology and the business context. Our AWS Well-Architected Review service covers all six pillars and produces a prioritised remediation roadmap, not just a list of findings.
Key takeaways
- AWS Trusted Advisor checks your account across five categories: cost, security, fault tolerance, performance, and service limits
- The full check library (300+ checks) requires a Business or Enterprise support plan
- Automate findings with EventBridge and Lambda instead of checking the console manually
- Pair Trusted Advisor with AWS Config for real-time enforcement and the Well-Architected Tool for architectural assessment
- For multi-account environments, use the organisational view to aggregate findings and track remediation progress
- Trusted Advisor is a starting point for cloud governance, not a complete solution on its own
Want help acting on your Trusted Advisor findings or running a full Well-Architected Review? Book a free consultation and we will assess your AWS account together.
Related service
Need hands-on help? Our certified engineers deliver this as a managed service.
AWS Well-Architected Review