1. Overview (Project Background)

S Company (Starbucks) prioritized reliably providing core services such as event systems where customer loyalty and traffic are extremely concentrated. While developing these core event systems, it was necessary to simultaneously solve two challenges: stable operation and cost efficiency of complex cloud infrastructure (EKS, RDS, etc.). Megazone Cloud combined infrastructure operation expertise (Managed Services) and AI-based development and operation automation (AIR DevOps) to support the success of this core project.

2. Challenge (Problem Definition)

• Peak time traffic instability: During frequency event periods, traffic surges suddenly, and infrastructure operation issues such as Aurora RDS load balancing issues and Redis Failover threatened service stability.

• Infrastructure operation complexity and cost increase: Continuous monitoring of various cloud components such as EKS, Cloudwatch, and VPC Flow Log, analysis of Cloudwatch cost increase causes, and cleanup of unused resources required complex infrastructure operation optimization efforts.

• Development-operations collaboration burden: During the core event development and deployment process, time delays occurred in technical review and work planning for DB performance (MySQL Parameter, Aurora Upgrade) and table column size changes.

3. Solution (Solution Approach)

Megazone Cloud provided solutions covering development, deployment, post-operation, and overall infrastructure optimization through the AIR DevOps methodology.

• High availability DB and infrastructure optimization support: Proactively reviewed Aurora MySQL 3 Upgrade methods and Blue/Green rollback strategies, and supported Redis Cluster Mode activation and automatic Reader Instance addition logic to prepare for peak time traffic.

• AI Ops-based cost and stability assurance: Provided analysis of Cloudwatch cost increase causes and cost reduction strategies, and achieved infrastructure cost efficiency through unused resource cleanup. Additionally, presented Cloudwatch standard alarm strategies and Whatap improvement plans to strengthen AI Ops-based monitoring systems.

• CI/CD and development environment consulting: Provided guidance on work strategies and time reduction methods for table column size changes, and proactively addressed EKS upgrade strategies and EKS Autoscaler issues to ensure stability of development and deployment environments.

4. Result (Results)

Through AIR DevOps support, we simultaneously achieved stable operation of the customer's event system and infrastructure cost efficiency.

• Core service operation stability assurance: Through vulnerability improvement and optimization of core infrastructure such as EKS and Aurora RDS (Redis Failover, EKS Security Group issues, etc.), we minimized service downtime during peak times.

• Development/operations efficiency improvement: Through proactive technical support such as table change work time reduction strategies and Aurora Upgrade rollback strategies, we shortened the lead time for development and DB-related tasks and improved operational efficiency.

• Cloud cost optimization: Through Cloudwatch cost reduction strategies and unused resource cleanup, we reduced operational costs and improved cloud infrastructure efficiency.