How Moka Reduced AWS Costs by ＄4500 per month in in 90 Days While Improving System Stability

Optimized AWS architecture and RDS performance to support high-concurrency HR workloads while significantly lowering long-term infrastructure costs.

Moka is a leading HR SaaS platform providing applicant tracking, recruitment automation, and talent management solutions for enterprise clients across Asia.
Its platform handles high-frequency operations such as candidate data processing, interview workflows, and real-time collaboration across distributed teams.
As customer adoption grew, Moka’s AWS infrastructure scaled rapidly—especially in database, compute, and cross-region traffic—leading to rising costs and performance bottlenecks during peak hiring periods.

Challenges

• RDS cost surge under high concurrency
  Recruitment peaks caused heavy database load, forcing over-provisioned instances.
• Inefficient EC2 resource allocation
  Static instance sizing failed to match fluctuating SaaS workloads.
• Excessive cross-region data transfer
  Multi-region architecture introduced unnecessary inter-region traffic costs.
• Lack of long-term cost optimization strategy
  No structured use of Savings Plans or Reserved Instances.
• Performance instability during peak hiring cycles
  Latency spikes impacted recruiter experience and system responsiveness.

Solution

• RDS performance tuning + right-sizing
  Optimized instance types, storage, and query performance to reduce over-provisioning.
• Dynamic EC2 optimization strategy
  Implemented right-sizing + auto-scaling aligned with real usage patterns.
• Cross-region traffic reduction
  Re-architected data flows to minimize unnecessary inter-region communication.
• Savings Plans implementation
  Introduced a structured commitment strategy based on usage baseline.
• Workload-aware architecture optimization
  Improved system stability during peak hiring periods with better resource distribution.

Results

Before vs After