---

name: performance description: Optimize application performance through caching strategies, load balancing, database scaling, and monitoring. Build systems handling thousands of concurrent users. sasmp_version: "2.0.0" bonded_agent: 05-caching-performance bond_type: PRIMARY_BOND

=== PRODUCTION-GRADE SKILL CONFIG (SASMP v2.0.0) ===

atomic_operations:

• BOTTLENECK_ANALYSIS
• CACHE_IMPLEMENTATION
• LOAD_BALANCING_CONFIG
• SCALING_STRATEGY

parameter_validation: query: type: string required: true minLength: 5 maxLength: 2000 current_rps: type: integer required: false description: "Current requests per second" target_latency_ms: type: integer required: false description: "Target P99 latency in milliseconds"

retry_logic: max_attempts: 3 backoff: exponential initial_delay_ms: 1000

logging_hooks: on_invoke: "skill.performance.invoked" on_success: "skill.performance.completed" on_error: "skill.performance.failed"

exit_codes: SUCCESS: 0 INVALID_INPUT: 1 METRICS_UNAVAILABLE: 2 OPTIMIZATION_FAILED: 3

Performance Optimization Skill

Bonded to: caching-performance-agent

---

Quick Start

# Invoke performance skill
"My API is slow, help me optimize it"
"Set up Redis caching for my application"
"Configure load balancing for high availability"

---

Instructions

1. Identify Bottlenecks: Profile application, analyze metrics
2. Choose Strategy: Select caching, scaling, or optimization approach
3. Implement Cache: Set up Redis/Memcached with appropriate pattern
4. Configure Scaling: Horizontal or vertical based on needs
5. Monitor Results: Set up APM and track improvements

---

Caching Patterns

Pattern	Use Case	Consistency	Complexity
Cache-Aside	Read-heavy, tolerates stale	Eventual	Low
Write-Through	Write-heavy, needs consistency	Strong	Medium
Write-Behind	High throughput writes	Eventual	High
Refresh-Ahead	Predictable access	Strong	Medium

---

Decision Tree

Performance Issue?
    │
    ├─→ High latency → Check database queries
    │     ├─→ Slow queries → Add indexes, optimize SQL
    │     └─→ Network → Add caching, reduce round-trips
    │
    ├─→ High CPU → Profile code
    │     ├─→ Algorithmic → Optimize algorithms
    │     └─→ Too much load → Scale horizontally
    │
    └─→ Memory issues → Analyze memory usage
          ├─→ Leaks → Find and fix leaks
          └─→ Large data → Implement pagination, streaming

---

Examples

Example 1: Redis Cache-Aside

import redis
import json
from functools import wraps

r = redis.Redis(host='localhost', port=6379, decode_responses=True)

def cache(ttl_seconds=3600):
    def decorator(func):
        @wraps(func)
        def wrapper(*args, **kwargs):
            cache_key = f"{func.__name__}:{args}:{kwargs}"
            cached = r.get(cache_key)
            if cached:
                return json.loads(cached)

            result = func(*args, **kwargs)
            r.setex(cache_key, ttl_seconds, json.dumps(result))
            return result
        return wrapper
    return decorator

@cache(ttl_seconds=300)
def get_expensive_data(user_id: str) -> dict:
    # Expensive database query or computation
    return {"user_id": user_id, "data": "..."}

Example 2: Connection Pooling

from sqlalchemy import create_engine

engine = create_engine(
    DATABASE_URL,
    pool_size=20,
    max_overflow=10,
    pool_timeout=30,
    pool_recycle=1800,
    pool_pre_ping=True
)

Example 3: Load Balancing (nginx)

upstream backend {
    least_conn;  # Least connections algorithm
    server backend1:8000 weight=3;
    server backend2:8000 weight=2;
    server backend3:8000 weight=1;

    keepalive 32;  # Connection pooling
}

server {
    location /api {
        proxy_pass http://backend;
        proxy_http_version 1.1;
        proxy_set_header Connection "";
    }
}

---

Performance Metrics

Latency Targets:
├── P50: < 50ms   (typical request)
├── P95: < 200ms  (most requests)
├── P99: < 500ms  (almost all)
└── P99.9: < 1s   (worst case)

Cache Metrics:
├── Hit Ratio: > 90% (ideal)
├── Eviction Rate: Low (stable)
└── Memory Usage: < 80% (headroom)

Database Metrics:
├── Connection Pool: < 80% utilized
├── Query Time: P95 < 100ms
└── Slow Queries: < 1% of total

---

Troubleshooting

Common Issues

Issue	Cause	Solution
High cache miss	Poor key design	Review key patterns
Cache stampede	Many misses at once	Use locks or jitter
Memory pressure	Over-caching	Set TTL, eviction policy
Connection timeout	Pool exhausted	Increase pool size

Debug Commands

# Redis stats
redis-cli INFO stats | grep -E 'keyspace|hits|misses'

# Memory usage
redis-cli INFO memory | grep used_memory_human

# Slow queries
redis-cli SLOWLOG GET 10

# Load test
wrk -t12 -c400 -d30s http://localhost/api

---

Test Template

# tests/test_performance.py
import pytest
import time

class TestCachePerformance:
    def test_cache_hit_is_fast(self, redis_client):
        # First call - cache miss
        start = time.time()
        result1 = get_cached_data("key1")
        cold_time = time.time() - start

        # Second call - cache hit
        start = time.time()
        result2 = get_cached_data("key1")
        warm_time = time.time() - start

        assert warm_time < cold_time * 0.1  # 10x faster
        assert result1 == result2

    def test_cache_ttl_expires(self, redis_client):
        set_cached_data("key", "value", ttl=1)
        time.sleep(2)
        assert get_cached_data("key") is None

---

Resources

• Redis Documentation
• High Performance Browser Networking
• Systems Performance (Brendan Gregg)