Mathews MwangiBack to projects
Distributed SystemsPlanned

KEDA Event-driven Autoscaling

Scale Kubernetes workers from zero based on Kafka lag and queue depth

Event-driven autoscaling with KEDA — worker Deployments that scale from zero based on Kafka consumer lag and task queue depth, demonstrating cost-efficient burst handling without over-provisioning.

View on GitHub

What it demonstrates

  • Scale-to-zero
  • Kafka scaler
  • Custom metrics
  • Burst handling

Tech stack

KubernetesKEDAKafkaHelm
All projects