What problem does it solve? Teams often treat performance and cloud spend as vague goals rather than measurable budgets, leading to overspending, premature scaling, and no clear signal of when capacity will run out. This Skill turns NFRs and baselines into a concrete performance and cost budget with targets, headroom, and guardrails. ## Core Features & Use Cases - Budget Definition: Derives per-Critical-User-Journey latency, throughput, and resource targets plus cost-per-unit budgets from NFRs and expected volume. - Cheap-Levers-First Plan: Orders optimization levers (caching, query/index tuning, cardinality, rightsizing) before any capacity or spend increase, with expected savings cited. - Headroom & Guardrails: Reports percent-to-ceiling and time-to-wall per dimension, plus a tiered guardrail hierarchy (soft alert, hard cap, circuit breaker). - Use Case: Given a wallet top-up service running at 100 RPS, produce a budget with p95 latency targets, cost-per-successful-top-up limits, connection-pool headroom, and a re-review trigger at 2x volume. ## Quick Start Ask the agent to set a performance and cost budget for your service, providing its critical user journeys, expected volume, and any baseline metrics.
