Case Study
Toko Digital Nusantara Kubernetes Multi-Region Autoscaling
Robust DevOps CI/CD pipeline and multi-region Kubernetes cluster autoscaling engineered to withstand 50x flash sale traffic spikes during national shopping festivals.
- Client
- PT Toko Digital Nusantara
- Industry
- E-Commerce Cloud Architecture Indonesia
- Service
- IT Support & Infrastructure
- Year
- 2024
- Pillar
- Innovation Development
Illustrative reference image; not documentation of the delivered project.
Starting point
Client Challenge
Sudden traffic surges during midnight flash sales overloaded container pods and caused sporadic checkout timeouts.
Kapitech response
Kapitech Solution
Implemented predictive horizontal pod autoscaling (KEDA), Redis cluster caching, Cloudflare Enterprise DDoS mitigation, and blue-green deployments.
Scope of work
Execution & Delivery
The recorded delivery scope for this project includes High-Concurrency Kubernetes Architecture, Automated GitLab CI/CD Pipeline, DDoS Protection & Cloudflare Edge Rules, and Real-Time Telemetry Dashboard. The technologies or production tools recorded for this engagement are Kubernetes, KEDA Autoscaler, Terraform, Redis Cluster, GitLab CI, and Datadog. The deliverables below make the documented scope easier to review.
Deliverables
- High-Concurrency Kubernetes Architecture
- Automated GitLab CI/CD Pipeline
- DDoS Protection & Cloudflare Edge Rules
- Real-Time Telemetry Dashboard
Results and indicators
Outcome & Impact
The impact indicators listed in this project record are Flash Sale Peak Requests, Checkout Success Rate, and Deployment Time. Available values are shown below so the reported outcomes can be reviewed in context.
- Flash Sale Peak Requests
- 850K Req/Sec
- Checkout Success Rate
- 99.98%
- Deployment Time
- 4 Mins vs 45 Mins
Technologies and tools
- Kubernetes
- KEDA Autoscaler
- Terraform
- Redis Cluster
- GitLab CI
- Datadog