Capacity Cost Model

capacity-cost-model

Turns "how much will this cost to run?" into a model a reviewer can re-run and finance can trust — sized from demand, expressed in cost per unit, and honest about the line items everyone forgets until the invoice. Distinct from business-case (which decides whether to invest, across options) and nfr-spec (which sets the throughput/availability targets); this models the resources and dollars to meet a target you've already chosen.

How to respond

1. Start from the demand model, never from instance types. The first question is "how much of what?" — the driver (requests/sec, tenants, GB stored, events/day, concurrent users), its current value, its growth curve, and the peak-to-average ratio. Ask at most 2 questions, spent on the driver's current volume + source and the peak ratio (the spikiness that decides static-vs-autoscale). Sizing to a round instance count instead of to demand is how clusters end up 80% idle or 200% over.

2. Express everything as unit economics. Cost per request / per tenant / per 1k events / per GB-month — because that's what scales predictably, what finance budgets in, and what reveals whether the architecture gets cheaper or more expensive per unit as it grows. ✅ "$0.012 per active tenant per day, dominated by the per-tenant search index" — ❌ "about $4k/month" (a number with no denominator can't be reasoned about at 10×).

3. Model the components against the cost catalog (reference.md) — compute, storage, network egress (the forgotten heavyweight), managed-service tiers, logging/observability (routinely a top-3 surprise), backups/DR, data transfer cross-AZ/region, and non-production environments (dev/staging/idle often 30–50% of the bill). Each line: the quantity from the demand model × the unit price, with the price's source/date tagged. Missing the egress and observability lines is the single most common way a model lands 40% low.

4. Make peak-vs-average and headroom explicit decisions, not defaults. State both the average load (what you bill for) and the peak (what you must serve), and the chosen target utilization + buffer with its reasoning: ✅ "size to peak × 1.3 for failover + spike; accept ~55% average utilization because the spike is revenue-critical". Over-provisioning burns money; under-provisioning is an incident — the headroom number is where that tradeoff is decided, so it's named, not buried in a rounded-up instance count.

5. Project at 1× / 3× / 10× and find the scaling cliffs. Cost rarely scales linearly — name where it jumps: the next instance/managed-service tier, cross-AZ egress kicking in, a license seat band, single-node → sharded, a free-tier ceiling. ✅ "linear to ~5×, then the search tier forces a sharded cluster (+$3k/mo step) at ~6×". A model that's a single multiplication hides the cliff that wrecks the budget.

6. Rank the cost drivers and use ranges. The 2–3 line items that dominate get named (optimization targets them, not the rounding error). Outputs are ranges with a stated lean (low/expected/high), never false precision — three estimated inputs don't produce "$4,271/mo". Tag input confidence like business-case does.

7. Set the cost guardrails. Budget alert thresholds, and the unit-cost regression check: cost-per-unit should hold flat or fall as volume grows; if the model shows it rising with scale, that's a flagged architectural problem, not a footnote. (Mirrors success-metrics' guardrail discipline; the alerting itself is an nfr-spec cost target.)

8. Emit with templates/cost-model.md in one message: demand model, component table with unit prices, unit economics, total at average + peak, growth scenarios with cliffs, cost-driver ranking, the forgotten-costs checklist (each ticked present or n/a), and guardrails. State the assumptions that most move the total at the top.

Useful references in this skill

• reference.md — the cost-component catalog, the forgotten-costs checklist, headroom guidance, and common scaling cliffs
• templates/cost-model.md — the model skeleton with demand, component, and growth-scenario tables

Quality bar

• Sized from a demand driver with a growth curve, not a chosen instance count. The driver and its source are stated.
• Every cost is a unit cost × a quantity, with the unit price's source and date tagged. A bare monthly total fails.
• Egress, observability, and non-prod lines are present (or explicitly n/a). Their absence is the signature of a model that's 40% low.
• Headroom and peak-vs-average are explicit decisions with reasoning, not absorbed into a rounded number.
• Scaling cliffs are named in the 1×/3×/10× projection. A linear extrapolation that hides a tier jump is a budget trap.
• Outputs are ranges with a lean; the unit-cost-at-scale trend is checked. Rising per-unit cost with growth is flagged as an architecture issue.

When to use this skill

• ✅ Estimating run cost for a new service/feature before launch
• ✅ Sizing capacity for a launch or a known growth event
• ✅ "What will this cost at 10× / when we hit N tenants?"
• ✅ Finding what's driving an existing cloud bill and where the next cliff is

When NOT to use this skill

• ❌ Deciding whether to build/buy/invest at all — that's business-case (this can feed its run-cost line)
• ❌ Setting the performance/availability targets — that's nfr-spec (this sizes to meet them)
• ❌ The migration/rollout sequence — that's migration-plan
• ❌ A fixed-price SaaS seat with no scaling math — just read the price sheet

Anti-patterns to avoid

• ❌ Sizing from instance types. "Let's run 3 large nodes" before knowing the load — that's a guess wearing a SKU, idle or overwhelmed by luck.
• ❌ The denominator-free total. "$4k/month" with no per-unit cost can't be projected, compared, or optimized.
• ❌ Forgetting egress and observability. The two line items that turn a tidy estimate into a 40%-low one; the checklist exists because they're forgotten every time.
• ❌ Ignoring non-prod. Three idle staging environments and a never-scaled-down dev cluster are real money the model pretends doesn't exist.
• ❌ Linear extrapolation past a cliff. "10× the load, 10× the cost" right up until the sharded-cluster step doubles it. Name the cliff.
• ❌ False precision. "$4,271.40/mo" from estimated inputs reads as certainty the model doesn't have. Ranges with a lean.
• ❌ Missing the rising-unit-cost flag. A model where cost-per-tenant climbs with scale is describing an architecture that gets worse the more it succeeds — that's the headline, not a row.