Scalability Advisor

Purpose

Assess system scalability: identify bottlenecks, recommend scaling strategies per container, and define capacity planning guidelines.

Input: Container diagram, NFRs (throughput, latency, users), data flow analysis (optional)
Output: Scalability Assessment with bottleneck analysis, scaling recommendations, and capacity model Examples: See examples.md for worked input/output.

Workflow

Step 1: Extract scalability requirements

From NFRs, define:

Metrics: concurrent users, requests/sec, data volume, growth rate.

Step 2: Analyze each container

For each container:

Identify: CPU-bound, memory-bound, I/O-bound, connection-pool limits.

Step 3: Identify system bottlenecks

Rank bottlenecks:

Common: database connections, single instance, sync external calls, large payloads.

Step 4: Recommend scaling strategies

Per container:

Step 5: Build capacity model

Simple capacity estimate:

Peak RPS = users × actions_per_session / session_duration
Required instances = Peak RPS / RPS_per_instance

Document assumptions explicitly.

Step 6: Define load testing plan

Step 7: Validate

Run Validation checklist.

Decision Rules

Validation

• Scalability NFRs extracted or assumptions stated
• Every container analyzed for bottleneck risk
• Top 3 bottlenecks ranked with rationale
• Scaling strategy per high-risk container
• Capacity model with explicit assumptions
• Load testing scenarios defined
• Premature optimization flagged if scale is small
• Trade-offs documented for each strategy

Anti-patterns

• Scale everything — horizontal scaling stateful DB without plan.
• Cache as fix-all — caching without invalidation strategy.
• Ignore externals — not modeling third-party rate limits.
• No load testing — recommendations without validation plan.
• Infinite scale claims — "handles any load" without evidence.

Best Practices

• Start with measured baseline, not guesses.
• Scale the bottleneck first (Amdahl's law).
• Prefer stateless containers for horizontal scaling.
• Plan scale-down as well as scale-up (cost).
• Align with data-flow-analyzer for data layer scaling.

Output Structure

# Scalability Assessment: [System Name]

## Requirements
| Metric | Target | Peak | Assumptions |
|--------|--------|------|-------------|

## Container Analysis
| Container | Stateless | Risk | Limit |
|-----------|-----------|------|-------|

## Bottlenecks
| Rank | Component | Constraint | At-risk metric |
|------|-----------|------------|----------------|

## Recommendations
| Container | Strategy | Rationale | Trade-off |
|-----------|----------|-----------|-----------|

## Capacity Model
[Calculations with assumptions]

## Load Test Plan
| Scenario | Load | Pass criteria |
|----------|------|---------------|

## Risks
| Risk | Mitigation |
|------|------------|

Next Skills