Advanced Settings

This section covers advanced configuration options for fine-tuning arb-assist performance and behavior.

Core Advanced Settings

update_interval

Type: integer (milliseconds) Default: 10000 (10 seconds) Description: How often to regenerate configuration files

update_interval = 5000   # Fast updates (5 seconds)
update_interval = 10000  # Standard (10 seconds)
update_interval = 30000  # Conservative (30 seconds)

Trade-offs:

• Lower: Fresher data, higher CPU usage

• Higher: Less resource usage, delayed reactions

run_interval

Type: integer (milliseconds) Default: 60000 (1 minute) Description: Minimum time between trades on the same mint

run_interval = 30000   # Aggressive (30 seconds)
run_interval = 60000   # Standard (1 minute)
run_interval = 300000  # Conservative (5 minutes)

Purpose:

• Prevents over-trading

• Allows market to settle

• Reduces failed transactions

halflife

Type: integer (milliseconds) Default: 120000 (2 minutes) Description: Time for statistics to decay by 50%

halflife = 60000   # Fast decay (1 minute)
halflife = 120000  # Standard (2 minutes)
halflife = 300000  # Slow decay (5 minutes)

Effects on different metrics:

• Profit: Recent profits weighted more

• Volume: Old volume fades out

• Success rate: Adapts to current performance

Visual Example of Half-life Decay

Time     | Weight
---------|--------
Now      | 100%
1 min    | 50%
2 min    | 25%
3 min    | 12.5%
4 min    | 6.25%

Advanced Filtering

ignore_filters

Type: boolean Default: false Description: Bypass all filter thresholds

ignore_filters = false  # Apply filters normally
ignore_filters = true   # Use any mint with 2+ pools

Use cases:

• Testing configurations

• Maximum opportunity capture

• With Jito no_failure_mode

Complex Filter Expressions

Combine multiple criteria:

filter_thresholds = [
  {
    # Basic requirements
    min_profit = 1_000_000,
    min_roi = 1.5,
    min_txns = 5,
    
    # Volume requirements
    min_net_volume = 100_000_000,
    min_total_volume = 1_000_000_000,
    
    # Balance requirements
    min_imbalance_ratio = 0.1,  # At least 10% directional
    max_imbalance_ratio = 0.9,  # Not more than 90% one-sided
    
    # Market depth
    min_liquidity = 10_000_000_000,
    min_turnover = 0.5,         # Volume/Liquidity ratio
    
    # Volatility
    min_volatility = 0.001,     # Minimum price movement
    
    # Profit consistency
    min_profit_per_arb = 100_000,  # Average profit per trade
  }
]

ALUTs (Address Lookup Tables)

aluts_per_pool

Type: integer Default: 20 Description: Number of ALUTs to collect per pool

aluts_per_pool = 10   # Minimal
aluts_per_pool = 20   # Standard
aluts_per_pool = 50   # Comprehensive

aluts

Type: array Default: [] Description: Custom ALUTs to always include

aluts = [
  "4sKLJ1Qoudh8PJyqBeuKocYdsZvxTcRShUt9aKqwhgvC",  # Custom ALUT 1
  "7nYmUgQPvvKN9vnt3xHukiGcgaqvBPfmFfkJcJQHHbsD",  # Custom ALUT 2
]

ALUT Strategy

For NotArb:

[notarb]
max_lookup_tables = 10  # Maximum ALUTs per transaction

Optimization tips:

• More ALUTs = larger transactions

• Fewer ALUTs = may miss addresses

• Monitor which ALUTs are actually used

Logging and Monitoring

log_output

Type: boolean Default: false Description: Enable detailed logging

log_output = true  # Enable verbose logging

Logs include:

• Mint rankings

• Filter decisions

• Configuration updates

• Performance metrics

Custom Memos

Add identifying information to transactions:

filter_thresholds = [
  { memo = "conservative-v1", ... },
  { memo = "moderate-v2", ... },
  { memo = "aggressive-v3", ... },
]

Mode-Specific Advanced Settings

SMB Advanced Settings

[smb]
output = "smb-config"
max_retries = 0              # Retry failed transactions
enable_simple_send = false   # Use simple send mode
cetiloan = true             # Enable flash loans
merge_mints = true          # Allow mint merging

cetiloan: Flash loan integration

• Borrow funds for arbitrage

• No upfront capital needed

• Small fee for borrowing

merge_mints: Combine multiple mints

• More complex transactions

• Higher potential profit

• Increased gas usage

NotArb Advanced Settings

[notarb]
output = "notarb-config"
jvm_args = ["-server", "-Xmx8192m", "-XX:+UseG1GC"]
threads = 0                  # 0 = dynamic, -1 = unlimited
flash_loan = true
meteora_bin_limit = 20       # Max Meteora DLMM bins

JVM Optimization:

jvm_args = [
  "-server",                 # Server JVM
  "-Xmx8192m",              # 8GB heap
  "-XX:+UseG1GC",           # G1 garbage collector
  "-XX:MaxGCPauseMillis=50" # 50ms max GC pause
]

Thread Management:

threads = 0    # Dynamic thread pool (recommended)
threads = -1   # Thread per task (maximum performance)
threads = 16   # Fixed thread pool (predictable)

Network Optimization

Connection Pooling

For NotArb:

[notarb]
# Separate RPC endpoints for different tasks
token_accounts_checker = { 
  rpc_url = "https://dedicated-checker.rpc.com", 
  delay_seconds = 3 
}

blockhash_updater = { 
  rpc_url = "https://fast-blockhash.rpc.com", 
  delay_ms = 400 
}

account_size_loader = { 
  rpc_url = "https://bulk-data.rpc.com", 
  invalid_account_size = 200, 
  buffer_size = 1500 
}

Benefits:

• Distribute load across endpoints

• Optimize for specific tasks

• Prevent rate limiting

File Server Configuration

port = 8080  # Enable file server
port = 0     # Disable file server

Security for file server:

# UFW firewall rules
sudo ufw allow from 192.168.1.0/24 to any port 8080
sudo ufw deny 8080

Performance Tuning

Memory Management

Control memory usage:

1. Reduce mints_to_rank

2. Increase halflife (faster decay)

3. Limit aluts_per_pool

4. Increase update_interval

CPU Optimization

Reduce CPU load:

1. Simplify sort strategies

2. Use fewer filter thresholds

3. Enable filter_programs

4. Increase delays

Network Optimization

Reduce bandwidth:

1. Filter GRPC stream

2. Batch RPC requests

3. Use compression

4. Cache static data

Special Configurations

Market Maker Mode

Focus on providing liquidity:

# High turnover, balanced flow
intermint_sort_strategy = { metric = "turnover", direction = "descending" }

filter_thresholds = [{
  min_turnover = 2.0,
  min_imbalance_ratio = 0.4,
  max_imbalance_ratio = 0.6,
  min_liquidity = 100_000_000_000,
}]

Sniper Mode

Target new opportunities:

# Sort by newest pools
intermint_sort_strategy = { metric = "pool_age", direction = "ascending" }

# Fast updates
update_interval = 5000
halflife = 30000  # 30 second decay

# Aggressive execution
spam_levels = [{
  process_delay = 200,
  tx_count = 5,
}]

Research Mode

Collect maximum data:

# Track everything
mints_to_rank = 100
filter_programs = false
ignore_filters = true
log_output = true

# Slow updates to reduce load
update_interval = 60000

Experimental Features

Custom Market Loaders

For NotArb:

[notarb]
market_loader = "custom-loader-endpoint"
lookup_table_loader = "custom-alut-endpoint"

Advanced Metrics

Custom sorting metrics (future):

# Potential future metrics
metric = "sharpe_ratio"      # Risk-adjusted returns
metric = "kelly_criterion"   # Optimal position sizing
metric = "entropy"          # Market efficiency

Best Practices

Configuration Management

1. Version Control

   • Track all config changes

   • Use meaningful commit messages

   • Tag stable configurations

2. Testing Protocol

   • Test new settings with small amounts

   • Monitor for 24 hours minimum

   • Compare performance metrics

3. Documentation

   • Comment complex configurations

   • Document parameter choices

   • Keep change log

Performance Monitoring

Key metrics to track:

• CPU usage percentage

• Memory consumption

• Network bandwidth

• Success rate

• Average ROI

• Configuration generation time

Gradual Optimization

1. Start with defaults

2. Change one parameter at a time

3. Measure impact

4. Document results

5. Iterate based on data

Troubleshooting Advanced Issues

Memory Leaks

Symptoms:

• Increasing memory usage over time

• System slowdowns

• OOM errors

Solutions:

• Reduce mints_to_rank

• Decrease aluts_per_pool

• Increase halflife

• Restart periodically

CPU Spikes

Symptoms:

• 100% CPU usage

• Delayed config updates

• System unresponsive

Solutions:

• Increase update_interval

• Simplify sort strategies

• Enable filter_programs

• Upgrade hardware

Network Saturation

Symptoms:

• Dropped GRPC connections

• RPC timeouts

• Incomplete data

Solutions:

• Upgrade bandwidth

• Filter GRPC stream

• Use multiple endpoints

• Implement caching