Key Questions:
- "What types of resources are we allocating?"
- "How many tenants and applications?"
- "What's the allocation latency requirement?"
- "Do we need fairness or priority-based allocation?"
- "How do we handle resource contention?"

"I'll design a resource allocation system with:

1. Components:
   - Allocation manager (scheduling)
   - Resource tracker (state)
   - Quota enforcer (limits)
   - Metrics collector (usage)

2. Scheduling:
   - Fair share within priority levels
   - Preemption for critical tasks
   - Anti-starvation mechanisms

3. Storage:
   - Redis for fast state
   - PostgreSQL for history
   - Prometheus for metrics"

✓ Fair resource distribution
✓ Starvation prevention
✓ High utilization (>80%)
✓ Low latency (<100ms)
✓ Multi-tenant isolation
✓ Quota enforcement

"First-come-first-served is always fair"
→ Can cause starvation

"We don't need preemption"
→ Critical for priority tasks

"Fair share with aging prevents starvation"
→ Shows understanding of fairness

"Preemption enables priority guarantees"
→ Demonstrates production awareness