magnus919_agent-skills/capacity-and-cost-engineering/templates/tenant-capacity-model.md

Tenant Capacity and Unit-Cost Model

Fill this template when tenant demand or tier promises change the capacity and cost decision. Use measured distributions where available; mark assumptions and synthetic profiles clearly.

Service and promise

• Service/resource boundary: [fill: API, worker pool, database partition, storage, etc.]
• Decision: [fill: sizing, placement, quota, tier promise, or cost allocation decision]
• Tenant tiers or profiles in scope: [fill: names and why they are representative]
• Customer promises: [fill: contractual promises, product defaults, and operational goals separately]
• SLO/performance dependency: [fill: owner and relevant target; route SLO definition to SRE]

Tenant demand profiles

Profile or cohort	Tenant count/weight	Demand and mix	Burst/concurrency	Storage/background work	Evidence/confidence
[fill: ordinary pooled tenant]	[fill]	[fill]	[fill]	[fill]	[fill]
[fill: hot or bursty tenant]	[fill]	[fill]	[fill]	[fill]	[fill]
[fill: dedicated/silo tenant]	[fill]	[fill]	[fill]	[fill]	[fill]

• Distribution view: [fill: per-tenant time series, selected statistic(s), correlation/skew analysis]
• Sampling limits: [fill: omitted tenants, seasonality, new-tenant uncertainty, synthetic data]

Skew and hot-tenant analysis

• Resource/partition boundary: [fill: shard, key range, worker, zone, cache, queue, or other]
• Detection signal: [fill: tenant-level and shared-resource signal]
• Contention path: [fill: which tenants or tiers are affected and how]
• Response: [fill: admission, queue, scheduling, throttling, placement, or isolation behavior]
• Recovery/rebalance: [fill: backlog, movement, reconciliation, and verification]
• Security handoff: [fill: isolation and authorization evidence owned by secure-software-engineering]

Pooled, siloed, or hybrid comparison

Scenario	Pooled baseline/headroom	Siloed baseline/headroom	Hybrid choice	Evidence and tradeoff
Ordinary load	[fill]	[fill]	[fill]	[fill]
Hot tenant or burst	[fill]	[fill]	[fill]	[fill]
Failure-domain loss	[fill]	[fill]	[fill]	[fill]
Tenant growth/movement	[fill]	[fill]	[fill]	[fill]

• Headroom rationale: [fill: what it protects, lead time, observed saturation, and cost of idle capacity; no universal percentage]

Quota, admission, and fairness

• Quota scope and units: [fill: tenant/tier/resource and steady/burst units]
• Admission rule: [fill: accept, queue, prioritize, throttle, or reject and why]
• Excess response: [fill: status/backpressure/degradation and customer communication]
• Fairness policy: [fill: explicit policy for this promise/resource, not a universal ratio]
• Fairness evidence: [fill: protected-tenant outcomes, constrained-tenant outcome, workload mix, duration, owner]
• Review trigger: [fill: what observed change causes recalibration]
• Implementation handoff: [fill: platform owner; this record does not configure infrastructure]

Tenant-variable unit cost

• Period and scope: [fill: same period for cost and demand]
• Platform baseline cost: [fill: idle/minimum pool, control plane, shared redundancy]
• Variable cost pool: [fill: compute, storage, transfer, jobs, support, or other measured costs]
• Shared-cost allocation: [fill: measured use, reserved entitlement, capacity reservation, blended, or other rationale]
• Tenant demand unit: [fill: request/job/GB/concurrency unit and measurement source]

tenant cost = allocated baseline + measured variable cost
tenant unit cost = tenant cost / tenant demand units

Tenant/profile	Allocated baseline	Variable cost	Total cost	Demand units	Unit cost	Confidence/gap
[fill]	[fill]	[fill]	[fill]	[fill]	[fill]	[fill]

• Financial handoff: [fill: pricing, margin, or commercial decision routed to financial-modeling]

Representative load/soak evidence

• Test plan: [fill: link to templates/load-soak-test-plan.md]
• Profiles exercised concurrently: [fill: ordinary, hot, skewed, tier, placement]
• Environment parity: [fill: data volume, topology, partitions, dependencies]
• Observed tenant outcomes: [fill: latency, errors, throttles, admission, backlog]
• Observed shared outcomes: [fill: saturation, partition balance, recovery, cost inputs]
• Soak findings: [fill: leaks, queue growth, fairness drift, or other trends]
• Verdict: [fill: validated / model with gap / failed; list criteria]

Assumptions, ownership, and decision

• Assumptions and gaps: [fill]
• Capacity/model owner: [fill]
• Demand measurement owner: [fill: route instrumentation to product-analytics-and-measurement]
• SLO/reliability owner: [fill: route SLO and error budget to site-reliability-engineering]
• Decision owner and review date: [fill]
• Chosen option and rationale: [fill]