Rootconf Mini 2024 (on 22nd & 23rd Nov)

Introduction

As cloud adoption surges, optimizing costs is vital, especially for startups. This article shares Xflow’s successful strategies that slashed our AWS bill from ~$25,000 to ~$11,500, saving ~$12,500 monthly after adjusting for additional non-AWS costs, resulting in a savings of ~$150,000 annually!

Architecture Overview

To illustrate our cost-saving measures, we provide a brief overview of our engineering architecture. Our APIs leverage API Gateway and EKS clusters for scalable service delivery. Transactional data is managed using MySQL, while search queries are handled by OpenSearch. We utilize Debezium and Kafka Connect for change data capture, with Kafka for message storage and Flink for processing. For observability, we employ a combination of CloudWatch and New Relic.

Component Wise Savings

• CloudWatch + NewRelic : ~$3500
• RDS : ~$2900
• Kinesis Data Analytics : ~$3700
• EKS+EC2 : ~$2000
• Others : -$400
• Total : ~$12500

We’ll explore each component’s optimization techniques, discuss opportunities like MSK vs. SQS where we could have saved cost, but chose not to and why, and share ongoing cost-saving initiatives.

Key Takeaways

• Effective strategies for AWS cost optimization
• Insights for startups on infrastructure cost management
• Guidelines for continuous cost control

This talk will provide a comprehensive overview of our cost reduction journey at Xflow, alongside practical strategies for maintaining budget efficiency.

Target Audience

• Engineering Leaders
• Architects/Cloud Engineers/SREs