← Back to Case Studies

Executive Summary

Led complete AWS cloud migration for affiliate marketing platform in 6 months with zero downtime for revenue-generating systems. Migrated 100% of infrastructure from managed hosting across 4 data centers to AWS, while simultaneously modernizing observability, introducing DevSecOps practices, and establishing deployment automation.

Key Results:

• Zero downtime for revenue-generating systems handling $1B+ annual transactions
• 100% infrastructure migrated in 6 months
• 4 data centers → 1 cloud platform consolidation
• 400%+ deployment frequency increase through GitOps automation
• MTTR improved from days to <1 hour for critical systems

The Challenge

Business Context

Situation (2022):

• Managed hosting provider became performance and support bottleneck
• Uneven performance, slow support responses
• Infrastructure not scaling with business needs
• Risk of service disruption from vendor dependency

Executive Decision:

• Migrate everything to AWS
• Timeline: 6 months (aggressive)
• Constraint: Zero tolerance for revenue disruption
• Budget approved for migration consultants

Technical Scope

Infrastructure to Migrate:

• 100% of production systems across 4 data centers
• Mix of physical servers and VMware virtual machines
• Legacy SQL Server databases (2012 Standard Edition)
• Windows file clusters
• Web services and application servers
• Network infrastructure (firewalls, load balancers)

Scale:

• Production systems handling $1B+ annual transaction volume
• 24/7 operations with high availability requirements
• Legacy systems with 15+ years of accumulated complexity

Risk Profile:

• Revenue disruption unacceptable
• Customer trust at stake
• Competitive market sensitive to service quality
• Acquisition integration pressures

The Approach

Phase 1: Strategy & Planning

Migration Strategy Evaluation:

Option A: Lift-and-Shift

• Move infrastructure as-is to AWS
• Minimal code changes
• Fastest path to cloud
• Selected for most systems

Option B: Re-architecture

• Redesign for cloud-native patterns
• Microservices, containers, serverless
• Maximum cloud benefits
• Too slow for 6-month timeline

Hybrid Approach Selected:

• Lift-and-shift foundation
• Targeted modernization where high-value
• Phased approach enabling incremental improvement

Cost/Risk/Effort Assessment:

• Produced detailed analysis for leadership
• Mapped dependencies and migration waves
• Identified technical debt requiring remediation
• Established success criteria and risk thresholds

Consultant Onboarding:

• Engaged external migration specialists
• Provided architecture briefings on ShareASale systems
• Shared documentation of quirks and gotchas
• Established clear scope and deliverables

Phase 2: Observability Migration First

Why Observability First:

• Can’t manage what you can’t measure
• Establish visibility before making changes
• Enable rapid incident response
• Validate migration success with metrics

From LogicMonitor to DataDog + CloudWatch:

LogicMonitor Context:

• Tied to managed hosting vendor being replaced
• Adequate but not cloud-native
• Migration opportunity to upgrade tooling

DataDog Benefits:

• Awin Global standard (alignment with parent company)
• Superior cloud-native integrations
• Better APM and distributed tracing
• More powerful query and alerting

Migration Execution:

• Audited existing monitors and alerts
• Identified gaps in coverage
• Created new DataDog pipelines and dashboards
• Integrated OpsGenie for on-call management
• Established SLO-based monitoring

Severity Classifications:

• Critical (Severity 1): Revenue-impacting, 15-min response SLA
• High (Severity 2): Customer-facing issues, 1-hour response
• Medium (Severity 3): Internal systems, 4-hour response
• Low (Severity 4): Non-urgent, next business day

Result:

• Comprehensive observability before migration began
• Clear metrics for validating each migration wave
• Rapid incident detection and response capability
• Foundation for DevSecOps culture shift

Phase 3: Infrastructure Migration Waves

Wave 1: SQL Server → AWS RDS

Source System:

• SQL Server 2012 Standard Edition
• Aging hardware approaching capacity
• Limited memory and I/O
• Manual backup processes

Target Architecture:

• AWS RDS SQL Server 2019 Enterprise Edition
• Always On Availability Groups (high availability)
• Automated backups with point-in-time recovery
• Multi-AZ deployment for failover

Migration Process:

• Schema and data profiling
• Performance testing in staging
• Phased cutover with fallback plan
• Monitoring and validation

Wave 2: File Clusters → AWS FSx

Source System:

• Windows file clusters on VMware
• Shared storage for application assets
• Manual replication processes
• Limited disaster recovery

Target Architecture:

• AWS FSx for Windows File Server
• Multi-AZ deployment
• Automated backups
• Integration with Active Directory

Benefits:

• High-performance file storage
• Automatic scaling
• Built-in disaster recovery
• Simplified management

Wave 3: Web Services → EC2 + Auto Scaling

Source System:

• Physical and virtual web servers
• Manual scaling processes
• Limited fault tolerance

Target Architecture:

• EC2 instances with Auto Scaling Groups
• Application Load Balancers (ALB)
• Launch templates for consistency
• Auto Scaling policies based on metrics

Modernization:

• GitOps-based deployment workflows
• Immutable infrastructure patterns
• Blue/green deployment capability
• Automated testing in pipeline

Wave 4: Network Configuration

Challenges:

• Translate firewall rules to Security Groups
• Load balancer configuration migration
• Network segmentation design
• VPN connectivity to legacy systems during transition

Solution:

• VPC design with public/private subnets
• Security Groups replacing firewalls
• Network ACLs for additional protection
• Transit Gateway for hybrid connectivity

Phase 4: Cutover & Validation

Phased Cutover Strategy:

• Non-critical systems first
• Canary deployments for production
• 24/7 monitoring during transition
• Rapid rollback capability

Contingency Planning:

• Detailed rollback procedures
• Communication plans for stakeholders
• War room coordination during cutover
• Post-mortem process for issues

Validation:

• Synthetic monitoring for health checks
• Performance comparison vs. baseline
• Customer-facing testing
• Business metric validation (transaction volumes, etc.)

Results

Migration Outcomes

Execution:

• Zero downtime for revenue-generating systems
• 100% infrastructure migrated successfully in 6 months
• 4 data centers → 1 cloud platform (AWS us-east-1)
• All waves completed on or ahead of schedule

Operational Improvements:

• Removed dependency on dysfunctional managed services provider
• Gained direct infrastructure control
• Enabled rapid scaling and provisioning
• Simplified disaster recovery

Recognition:

• Featured in AWS case study (since removed from public site)
• Executive satisfaction with delivery
• Zero customer impact during migration
• Team morale boost from successful delivery

Cultural Transformation

DevSecOps Practices:

• Introduced “Shift Left” security principles
• Code review and automated testing
• Security scanning in CI/CD pipeline
• Compliance-as-code patterns

GitOps Workflows:

• Migrated from SVN to GitLab
• Git-based deployment automation
• Infrastructure as Code (Terraform concepts)
• Version-controlled configuration

Deployment Automation:

• 400%+ deployment frequency increase
• Manual deployments → GitLab CI automation
• Reduced human error through automation
• Faster time-to-market for features

Incident Response:

• MTTR: Days → <1 hour for critical systems
• MTTR: Hours → ~5 minutes for highest-priority systems
• Comprehensive monitoring and alerting
• Runbooks and escalation procedures

Lessons Learned

What Worked Well

Observability First:

• Having visibility before migration was critical
• Enabled rapid issue detection and resolution
• Validated migration success with metrics
• Built confidence in migration process

Consultant Partnership:

• External expertise accelerated execution
• Knowledge transfer improved internal capability
• Risk mitigation through experience
• But required strong internal coordination

Phased Approach:

• De-risked migration through incremental waves
• Enabled learning and course corrections
• Maintained business continuity
• Rollback capability provided safety net

What I’d Do Differently

More Aggressive Modernization:

• Lift-and-shift was safe but left technical debt
• Could have re-architected more systems
• Containers would have simplified operations
• But timeline pressure justified pragmatism

Earlier Automation:

• Some manual processes remained post-migration
• Should have automated more during migration
• Infrastructure as Code adoption could have been faster

Better Documentation:

• Under time pressure, documentation lagged
• Should have documented architecture decisions real-time
• Knowledge transfer could have been more systematic

What This Demonstrates

For Cloud Migration Roles:

• Complete AWS migration execution experience
• Zero-downtime migration methodology
• 6-month aggressive timeline delivered
• Risk management and mitigation

For SRE / Infrastructure Roles:

• High-availability architecture design
• Disaster recovery planning and execution
• Observability platform migration (DataDog, CloudWatch)
• Incident response process establishment

For DevOps / Platform Engineering:

• GitOps workflow implementation
• CI/CD pipeline establishment
• Deployment automation (400%+ frequency increase)
• Infrastructure modernization

For Technical Leadership:

• Multi-phase program execution
• Consultant management and coordination
• Stakeholder communication (C-level to engineers)
• Cultural transformation (DevSecOps adoption)

Technologies Used

AWS Services:

• RDS (SQL Server with Always On)
• FSx (Windows File Server)
• EC2 (Auto Scaling Groups, Launch Templates)
• VPC (networking, Security Groups, NACLs)
• Application Load Balancer
• CloudWatch (monitoring and logging)

Observability:

• DataDog (APM, logs, metrics, dashboards)
• CloudWatch (AWS-native monitoring)
• OpsGenie (incident management, on-call rotation)

Development & Deployment:

• GitLab (source control, CI/CD)
• GitLab CI (pipeline automation)
• PowerShell (automation scripts)
• Terraform concepts (infrastructure as code)

Migration Tools:

• AWS Server Migration Service
• Database migration utilities
• Custom scripts for configuration replication

Contact

Led AWS cloud migrations or considering one for your organization? Let’s discuss zero-downtime strategies, observability-first approaches, and cultural transformation.

Get in Touch: stevenleve.com/contact
LinkedIn: linkedin.com/in/steve-leve

← Back to Case Studies