scaling-plan-authoring

Scaling & Capacity Plan Authoring — Authors threshold-gated scaling strategies, capacity baselines, bottleneck analysis, auto-scaling policies, and tiered infrastructure cost models.

SKILL.md

GitHub

SKILL: Scaling & Capacity Plan Authoring

Domain: Performance Engineering, Capacity Planning, SDD Station 15 / 18

Activation triggers: scaling plan, capacity planning, traffic growth, horizontal scaling, bottleneck analysis, cloud infrastructure costs, SDD Station 15.


1. Role & Threshold Law

The Scaling Plan Authoring skill evaluates and plans system behavior under extreme concurrency and data volume.
It enforces the Scale Threshold Law:

  • The Scaling Plan is MANDATORY whenever:
    1. Traffic targets exceed 500 RPS or 10,000 DAU.
    2. Growth expectations anticipate > 10x scale within 12 months.
    3. Multi-region or multi-AZ active-active failover is required.
    4. The PRD contains explicit horizontal auto-scaling non-functional requirements.
  • For internal tools or targets below these thresholds, the document may record a formal threshold exemption.

2. Document Structure & Required Sections

Every generated Scaling Plan must follow .factory/specifications/templates/10_scaling_plan.template.md and contain:

  • Header Metadata: Document ID, Version, Status, Parent PRD/Architecture IDs, Lead Performance Engineer, Date.
  • Section 1: Applicability Threshold & Scale Trigger: Explicit justification of why the plan is required or exempt.
  • Section 2: Capacity Baseline & Growth Projections:
    • Table showing Baseline (Current/MVP), Stage 1 (10x), and Stage 2 (Enterprise Scale/50x) across DAU, Peak RPS, Data Volume, and WebSocket connections.
  • Section 3: Bottleneck Analysis by Architectural Layer:
    • In-depth risk, symptom, and mitigation analysis for:
      • Database & Persistence Tier (connection exhaustion, read replicas, partitioning).
      • Backend API & Compute Tier (CPU saturation, stateless containers, HPA).
      • Network & Ingestion Tier (bandwidth, CDN caching, rate limiting).
  • Section 4: Horizontal vs. Vertical Scaling Strategy:
    • Matrix detailing scaling triggers and instance limits per container component.
  • Section 5: Load Testing Benchmarks:
    • Concrete stress limits and degradation thresholds referenced from 09_testing_plan.md.
  • Section 6: Infrastructure Cost Projections by Scale Tier:
    • Monthly cloud spend estimates and cost drivers across Baseline, Growth, and Enterprise tiers.
  • Section 7: Signoff & Verification.

3. Validation Checklist (Quality Gate G11 & G15 Pre-Condition)

  • Clear threshold justification documented.
  • Bottleneck mitigations defined for database, compute, and network.
  • Auto-scaling triggers and container limits specified.
  • Tiered cloud infrastructure costs estimated.
  • Document saved to .factory/specifications/<project>/10_scaling_plan.md.