Skip to main content
The Recommendation Engine is the core of CloudThinker’s cost optimization system. It analyzes your infrastructure, identifies savings opportunities, and generates actionable recommendations with clear implementation paths.

​
Recommendation Attributes

Each recommendation includes comprehensive metadata to help you prioritize and implement:
AttributeDescription
TitleClear, actionable summary of the optimization
DescriptionDetailed explanation of the issue and solution
Potential SavingsEstimated monthly or annual savings (high precision)
Effort LevelImplementation complexity: Low, Medium, or High
Risk LevelPotential impact on workloads: Low, Medium, or High
SourceOrigin: Assessment, Conversation, Manual, or CloudKeepers
StatusCurrent state: Pending, In Progress, Implemented, or Ignored
VisibilityWorkflow state: Draft, Active, or Archived

​
Recommendation Lifecycle

1

Generated

Alex or CloudKeepers identifies an optimization opportunity and creates a recommendation in Draft status.
2

Active

Review and promote recommendations to Active to make them visible to your team and trackable.
3

In Progress

Mark recommendations as In Progress when implementation begins.
4

Implemented

After completion, mark as Implemented to track actual savings against projections.
5

Ignored

Dismiss recommendations that aren’t applicable with an optional reason.

​
Viewing Recommendations

​
Dashboard View

Access recommendations from the main dashboard to see:
  • High-priority recommendations by potential savings
  • Recommendations grouped by resource type
  • Implementation status overview
  • Total potential and implemented savings

​
Conversation-Based Discovery

Ask Alex to find and explain recommendations: 
# List top recommendations
@alex show top 10 cost recommendations by savings

# Filter by resource type
@alex what EC2 optimization opportunities exist?

# Filter by effort
@alex show low-effort recommendations with high savings

# Explain a specific recommendation
@alex explain why we should resize instance i-0abc123

​
Creating Recommendations

Recommendations are created through multiple sources:

​
1. Automated Analysis

CloudKeepers continuously monitors your infrastructure and generates recommendations based on:
  • Resource utilization patterns
  • Cost anomalies
  • Best practice violations
  • Spending trends

​
2. Agent Conversations

Ask Alex to analyze specific areas: 
@alex analyze our S3 storage costs and recommend optimizations

@alex review EC2 instances that have been idle for over 7 days

​
3. Well-Architected Assessments

Recommendations are generated during Well-Architected Framework assessments under the Cost Optimization pillar.

​
4. Manual Creation

Create recommendations manually for custom optimization opportunities:
  • Navigate to Recommendations in your workspace
  • Click “New Recommendation”
  • Fill in the recommendation details
  • Assign effort, risk, and savings estimates

​
Recommendation Categories

​
Compute Optimization

Identify instances with consistently low CPU/memory utilization that can be downsized:
  • EC2 instances under 20% average utilization
  • RDS instances with excess capacity
  • Lambda functions with over-provisioned memory
Recommendations for converting on-demand to reserved instances:
  • 1-year vs 3-year commitment analysis
  • Savings Plans coverage gaps
  • Reserved instance utilization optimization
Identify workloads suitable for Spot instances:
  • Fault-tolerant batch jobs
  • Development/test environments
  • Stateless applications

​
Storage Optimization

Automate storage tier transitions:
  • S3 Intelligent Tiering enablement
  • Glacier archive recommendations
  • Infrequently accessed data identification
Identify and cleanup:
  • Unattached EBS volumes
  • Orphaned snapshots
  • Empty S3 buckets
  • Unused EFS file systems

​
Database Optimization

  • Over-provisioned RDS instances
  • DynamoDB capacity mode recommendations
  • ElastiCache node optimization
  • Index recommendations for query performance
  • Query optimization suggestions
  • Read replica opportunities

​
Network Optimization

  • Cross-region transfer optimization
  • NAT Gateway efficiency
  • VPC endpoint recommendations
  • Idle load balancer detection
  • ALB to NLB migration opportunities
  • Cross-zone load balancing optimization

​
Discussion & Collaboration

Each recommendation includes a discussion thread for team collaboration:
  • Comments: Add context, questions, or implementation notes
  • Mentions: Tag team members with @username
  • Attachments: Link related documents or tickets
  • Audit Trail: Track all changes and status updates

​
Implementing Recommendations

​
With Approval

For recommendations requiring infrastructure changes:
  1. Review the recommendation details
  2. Request implementation from Alex
  3. Review the proposed changes
  4. Approve execution in the Approval Center
  5. Monitor implementation progress
@alex implement recommendation for resizing instance i-0abc123

# Alex will show the proposed changes and request approval

​
Manual Implementation

For recommendations you want to implement yourself:
  1. Mark as “In Progress”
  2. Follow the implementation steps provided
  3. Use the discussion thread for questions
  4. Mark as “Implemented” when complete
  5. Add actual savings for tracking

​
Tracking & Reporting

​
Savings Dashboard

Track optimization progress with:
  • Total potential savings identified
  • Savings implemented vs. available
  • Savings trend over time
  • Implementation velocity

​
Export & Integration

  • Export recommendations to CSV/Excel
  • Create Jira tickets from recommendations
  • Sync with external tracking systems via webhooks
  • Generate executive reports

Configure Webhooks

Send recommendation updates to external systems