Autoscaling Controller  ·  View 22 of 37  ·  4 · Runtime

Scale-Out

A sustained backlog breach, through the warm pool and past it, to registered capacity.

Editable source SVG draw.io All views
Work Queue Decision Engine Desired State Reconciler Warm Pool VMSS Load Balancer 1. backlog age rising 2. interval 1: breach, hold 3. interval 2: sustained 4. desired 40 -> 72 5. read desired 72 6. claim 8 warm members 7. attach · under 10 s 8. provision 24 cold 9. boot, warm, health 10. registered: counts now 11. scale-in suppressed Scale-Out — Sustained Breach to Serving Capacity p95 to serving capacity: 60 s for the warm-pool-backed share, 240 s for the cold share. The warm pool exists to cover the first of those windows, not the second. v 1.0 · owner Data & AI Global Practice · date 2026-09

Read this for

  • Interval 1 sees the breach and does nothing. The sustain window is why the controller does not chase a spike.
  • The warm pool covers the first 8 instances in under 10 s; the remaining 24 are cold and take the full provision-to-registered path.
  • Scale-in is suppressed from the moment the delta is applied until the last new instance is serving.

Targets

  • p95 sustained breach to serving: 60 s warm-pool-backed, 240 s cold.
  • Warm-pool attach under 10 s; a miss is recorded as a latency penalty against the target, not as a failure.

Assumptions

  • 40 → 72 is inside the +100%/60 s growth cap; a larger jump would be applied over two windows and journalled as rate-limited.
◀  Stability and Oscillation Control All views Scale-In  ▶
Solution Architecture v1.0 · Data & AI Global Practice · 2026-09 · 37 views · Azure and open source22-scale-out-sequence