BMAD Performance Optimization Skill
When to Invoke
Trigger this skill when the user:
- Reports latency, throughput, or resource regressions.
- Requests load/performance testing guidance or results interpretation.
- Needs to set or validate performance budgets and SLAs.
- Wants to plan scaling strategies ahead of a launch or marketing event.
- Asks how to tune code, queries, caching, or infrastructure for speed.
If the user only needs to implement a specific optimization already defined, delegate to bmad-development-execution.
Mission
Deliver actionable insights, testing strategies, and prioritized optimizations that keep the product within agreed performance budgets while balancing cost and complexity.
Inputs Required
- Current architecture diagrams and deployment topology.
- Observability data: metrics dashboards, traces, profiling dumps, load test reports.
- Performance requirements (SLAs/SLOs, budgets, target response times).
- Workload assumptions and peak usage scenarios.
Gather missing telemetry by coordinating with bmad-observability-readiness if instrumentation is lacking.
Outputs
- Performance brief summarizing current state, key bottlenecks, and risks.
- Benchmark and load test plan aligning tools, scenarios, and success criteria.
- Optimization backlog ranked by impact vs. effort with owner and verification plan.
- Updated performance budget recommendations or SLO adjustments when necessary.
Process
- Validate inputs and ensure instrumentation coverage. Escalate gaps to observability skill.
- Analyze telemetry to pinpoint hotspots (CPU, memory, I/O, DB, network, frontend paint times).
- Assess architecture decisions for scalability (caching, asynchronous workflows, data partitioning).
- Define performance goals and acceptance thresholds with stakeholders.
- Create load/benchmark plans covering baseline, stress, soak, and spike scenarios.
- Recommend optimizations across code, database, infrastructure, and CDN layers.
- Produce backlog with measurable acceptance criteria and regression safeguards.
Quality Gates
- Recommendations trace back to observed data or projected workloads.
- Each backlog item includes measurement approach (before/after metrics).
- Performance budgets and SLAs updated or reaffirmed.
- Risks communicated when goals require major architectural change.
Error Handling
- If telemetry contradicts assumptions, schedule hypothesis-driven experiments rather than guessing.
- Flag when performance targets are unrealistic within constraints; propose trade-offs.
- When required tooling is unavailable, document blockers and coordinate with observability & dev skills.
1---2name: bmad-performance-optimization3description: Diagnoses bottlenecks and designs performance optimization plans.4---5
6# BMAD Performance Optimization Skill
7
8## When to Invoke
9
10Trigger this skill when the user:
11- Reports latency, throughput, or resource regressions.
12- Requests load/performance testing guidance or results interpretation.
13- Needs to set or validate performance budgets and SLAs.
14- Wants to plan scaling strategies ahead of a launch or marketing event.
15- Asks how to tune code, queries, caching, or infrastructure for speed.
16
17If the user only needs to implement a specific optimization already defined, delegate to `bmad-development-execution`.
18
19## Mission
20
21Deliver actionable insights, testing strategies, and prioritized optimizations that keep the product within agreed performance budgets while balancing cost and complexity.
22
23## Inputs Required
24
25- Current architecture diagrams and deployment topology.
26- Observability data: metrics dashboards, traces, profiling dumps, load test reports.
27- Performance requirements (SLAs/SLOs, budgets, target response times).
28- Workload assumptions and peak usage scenarios.
29
30Gather missing telemetry by coordinating with `bmad-observability-readiness` if instrumentation is lacking.
31
32## Outputs
33
34- **Performance brief** summarizing current state, key bottlenecks, and risks.
35- **Benchmark and load test plan** aligning tools, scenarios, and success criteria.
36- **Optimization backlog** ranked by impact vs. effort with owner and verification plan.
37- Updated performance budget recommendations or SLO adjustments when necessary.
38
39## Process
40
411. Validate inputs and ensure instrumentation coverage. Escalate gaps to observability skill.
422. Analyze telemetry to pinpoint hotspots (CPU, memory, I/O, DB, network, frontend paint times).
433. Assess architecture decisions for scalability (caching, asynchronous workflows, data partitioning).
444. Define performance goals and acceptance thresholds with stakeholders.
455. Create load/benchmark plans covering baseline, stress, soak, and spike scenarios.
466. Recommend optimizations across code, database, infrastructure, and CDN layers.
477. Produce backlog with measurable acceptance criteria and regression safeguards.
48
49## Quality Gates
50
51- Recommendations trace back to observed data or projected workloads.
52- Each backlog item includes measurement approach (before/after metrics).
53- Performance budgets and SLAs updated or reaffirmed.
54- Risks communicated when goals require major architectural change.
55
56## Error Handling
57
58- If telemetry contradicts assumptions, schedule hypothesis-driven experiments rather than guessing.
59- Flag when performance targets are unrealistic within constraints; propose trade-offs.
60- When required tooling is unavailable, document blockers and coordinate with observability & dev skills.
majiayu000