{
  "id": 10409121,
  "title": "Uber Separates Scaling Intent From Execution on Kubernetes Platform",
  "url": "https://urgent.news/2026/09/28/uber-separates-scaling-intent-from-execution-on-kubernetes-platform",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-09-28T09:00:00.000Z",
  "source": {
    "name": "InfoQ",
    "slug": "infoq",
    "url": "https://www.infoq.com/news/2026/09/uber-kubernetes-scaling/"
  },
  "original_language": "en",
  "account": "Uber has unveiled its ServiceScale controller, enabling multiple orchestrators to manage scaling of the same Kubernetes workloads safely. This controller, detailed in a blog post by software engineers Egor Grishechko and Srikar Paruchuru, separates scaling intent from execution. This separation allows for regional failover without maintaining idle capacity. Uber's Container Platform team oversees over 100 compute clusters, handling 4,000 services on 3 million cores with 1.5 million daily pod launches. The team's internal platform, Up, serves as a federation layer for the Kubernetes fleet. Services are deployed via Up, with scaling expectations set, then reconciled into Kubernetes primitives by the Uber Deployment Controller (UDC). Prior to this, Uber reserved idle capacity in all data centers for regional failovers. This led to the introduction of a new custom resource definition, ServiceScale, and the Service Scale Controller (SSC). Each orchestrator can express its scaling desire through ServiceScale, and SSC reconciles this intent into Kubernetes objects. Keeping the model simple, they avoided an external database or separate coordination service. The new system improves inspection and failback, as both steady-state and temporary failover information are saved in the CRD spec. Uber learned that stale informer caches and multi-writer systems presented challenges. They implemented read-your-own-write consistency guards and observability to detect metadata-spec drift, leading to a long-term fix in the scaling path. The rollout reduced steady-state provisioning from 2x to 1.3x and eliminated over a million CPU cores.",
  "summary": "Uber has published a detailed account of its new ServiceScale controller, which allows multiple orchestrators to safely manage the scaling of the same Kubernetes workloads. The blog post, written by senior software engineers Egor Grishechko and Srikar Paruchuru, describes how the company separated scaling intent from execution to support regional failover without carrying reserved idle 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."
}