Performance Tuning

Optimize your node for maximum rewards.

---

Key Metrics

Metric	Target	Impact
Latency (p99)	< 10ms	Win rate
Cache Hit Rate	> 90%	Speed
Uptime	> 99.9%	Selection
Accuracy	100%	Rewards

---

Database Tuning

[database]
memory_limit = "80%"
threads = 0  # Auto-detect
cache_size_mb = 4096
wal_mode = true

DuckDB Optimizations

-- Increase memory
PRAGMA memory_limit='64GB';

-- Use parallel execution
PRAGMA threads=16;

-- Enable statistics
PRAGMA enable_progress_bar;

---

Cache Optimization

[performance]
cache_ttl_seconds = 5
max_concurrent_queries = 100

[specialization]
cache_capacity_gb = 32
eviction_policy = "lru"

Hot Data Strategy

1. Monitor access patterns
2. Pre-warm cache on startup
3. Tune TTL based on data freshness needs

---

Network Optimization

# Increase socket buffers
sudo sysctl -w net.core.rmem_max=134217728
sudo sysctl -w net.core.wmem_max=134217728

# Enable TCP fast open
sudo sysctl -w net.ipv4.tcp_fastopen=3

---

System Tuning

# Disable swap
sudo swapoff -a

# Set CPU governor
sudo cpupower frequency-set -g performance

# Increase file limits
ulimit -n 1000000

---

Monitoring

# Watch key metrics
streamsync metrics watch

# Export to Prometheus
curl localhost:9090/metrics

---

Benchmarking

# Run local benchmark
streamsync benchmark --queries 10000

# Compare with network
streamsync benchmark --network