prompt
FreeAI-powered performance diagnosis and optimization prompt
FreeFree tier
Inputs: textOutputs: text
About prompt
A specialized prompt template from the ai-boost/awesome-prompts repository that configures an AI to act as a seasoned performance engineering expert. It guides the AI through a structured five-step process to diagnose performance bottlenecks (CPU, I/O, memory) in web applications, APIs, and data pipelines, and to produce a prioritized optimization plan with concrete code examples and validation methodology. The prompt includes detailed roles, context, input handling (symptoms, tech stack, metrics), task decomposition, output specifications, and quality criteria to ensure evidence-based, actionable recommendations.
Key Features
Structured 5-step performance diagnosis process (baseline, bottleneck analysis, impact quantification, implementation plan, validation)
Role-based system prompt with 12+ years expert persona
Handles diverse symptoms: slow endpoints, high CPU, memory growth, lock contention
Recommends profiling tools (py-spy, pprof, async-profiler, Chrome DevTools) and APM platforms (Datadog, New Relic, Jaeger)
Produces prioritized optimization list sorted by impact/effort ratio
Includes concrete code or query examples and a measurement plan
Avoids generic advice like just add caching without identifying what to cache
Pros & Cons
Pros
- Provides a rigorous, evidence-based methodology for performance problem-solving
- Quantifies expected improvements and scores optimizations by impact vs. complexity
- Includes validation approach to prevent regressions
- Free and open-source, hosted on GitHub with 8.5k stars
- Works with any LLM that accepts system prompts (e.g., ChatGPT, Claude)
Cons
- Requires manual input of performance symptoms and tech stack details
- Output is text-based, not integrated with monitoring tools or dashboards
- Quality of diagnosis depends on user's ability to gather accurate metrics and profiling data
- May need iterative refinement for complex or unusual bottlenecks
Best For
Debugging slow API endpoints in productionOptimizing database query performance (N+1, missing indexes, lock contention)Reducing CPU or memory usage in web applicationsImproving latency for high-traffic services (100-1000 req/min)Building performance testing and regression monitoring strategies
FAQ
What inputs does this prompt require?
The prompt asks for an observed performance symptom (e.g., slow endpoint, high CPU), the technology stack (language, framework, database), and any metrics already collected (response times, CPU%, query times). If not provided, it infers moderate traffic (100-1000 req/min) and standard cloud infrastructure.
What kind of output does this prompt produce?
It produces a diagnosis followed by a prioritized optimization list with implementation guidance, typically 400-700 words. The output includes a root cause hypothesis, an optimization list sorted by impact/effort ratio, at least one concrete code or query example, and a measurement plan to verify improvements.
Is this prompt free to use?
Yes, the prompt is part of the open-source ai-boost/awesome-prompts repository on GitHub and is completely free to use. You can copy the text and use it with any compatible AI model.
Who created this prompt?
The prompt is contributed to the ai-boost/awesome-prompts community repository on GitHub. The specific author is not credited in the file itself.
Can I modify this prompt for my specific needs?
Yes, since it is open-source, you are free to modify it. The repository encourages customization and improvement.