{
  "id": 765404,
  "title": "KEDA 3.0 Scale-to-Zero: How We Cut Intermittent Kubernetes Workload Costs to Almost Nothing",
  "url": "https://urgent.news/2026/08/13/keda-3-0-scale-to-zero-how-we-cut-intermittent-kubernetes-workload",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-08-13T12:30:25.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/muskan_bandta/keda-30-scale-to-zero-how-we-cut-intermittent-kubernetes-workload-costs-to-almost-nothing-2egn"
  },
  "original_language": "en",
  "account": "KEDA 3.0 just released with a game-changing feature: event-driven autoscaling that includes over 80 event sources like Kafka and RabbitMQ, and supports scale-to-zero. This means workloads that sit idle most of the day and spike when work arrives can now pay for capacity only when they need it, rather than always being over-provisioned. In our case, intermittent workloads that were idle 80%+ of the time saw significant cost savings after switching to this new autoscaling pattern. However, there are some considerations to keep in mind. One of the main drawbacks is the cold start delay that occurs when scaling from zero - the first event waits for a pod to schedule, pull its image, and start. For lightweight consumers, this delay is usually a few seconds, but for heavier images or slow-warming JVMs, the delay can be much longer. Another factor to consider is that while KEDA scales pods to zero, the node still needs to exist. If scaling to zero leaves a node empty, you only capture the savings when the cluster autoscaler or Karpenter actually removes that node. Lastly, the cooldownPeriod setting is a cost lever rather than a default - if set too short, you'll experience frequent scaling and incur repeated cold starts. If set too long, you'll keep pods alive even after the last event. Tuning this per workload is crucial to optimizing costs. Overall, KEDA 3.0's scale-to-zero feature is a powerful tool for cost optimization of intermittent workloads, but it requires careful consideration and tuning to fully realize its benefits.",
  "summary": "KEDA 3.0 just landed, and the headline feature is the one I care about most as someone who watches a cloud bill: event-driven autoscaling now covers 80+ event sources (Kafka, RabbitMQ, and a long list more) with proper scale-to-zero. If you run workloads that sit idle most of the day and spike when work arrives, this is the difference between paying for capacity you use and paying for capacity…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}