6.8 Monitoring & Operations

Once live, you need visibility, not blind faith.

Typical dashboard:

• PnL

  • Realized PnL (daily / weekly / monthly).

  • Unrealized PnL on open routes (if any).

• Trade stats

  • Number of trades.

  • Win rate.

  • Average / median profit per trade.

  • Edge decay (expected vs realized profit per trade).

• Risk usage

  • Current drawdown vs daily cap.

  • Max trade size vs configured max.

  • Exposure per strategy type.

• Execution health

  • Error counts (failed tx, RPC issues).

  • Average latency from opportunity detection to tx sent.

  • MEV slippage incidents (where fills deviated hard from expected).

Controls you must have:

• Pause / stop

  • Immediate halt button.

  • Optional auto-halt triggers (e.g. too many infra errors in a window).

• Config change

  • Change thresholds (min profit, risk per trade) without redeploying.

  • Change strategy mix on the fly (enable/disable classes).

If you can’t see this data and can’t stop the system instantly, you’re not running a bot, you’re gambling.