what is trading-fleet-infrastructure fr?

This project is a system for running many automated trading strategies at once, using fake money against real market prices, on a single always-on cloud server. Fifteen strategies covering stocks, options, and crypto run continuously, each as its own independent program with its own simulated brokerage account, its own risk limits, and its own alert channel. The author is explicit that this is not about proving any one strategy is profitable. With only weeks of live data, any results would just be noise, so the project makes no performance claims. The real subject is the operational engineering behind keeping many unattended trading programs running reliably. Each strategy runs as a supervised background service on the server, automatically restarting if it crashes and capped in how much memory it can use, with its logs rotated and compressed so they do not pile up. A shared component calculates the day's overall market condition and shares it with strategies that adjust their behavior based on it. Another shared component, the risk engine, checks every account's value daily and enforces a hard stop, called a kill line, that flags a strategy for shutdown if its losses go too far. Every night, a separate job pulls every trade from all accounts, matches them into profit and loss figures per strategy, and rebuilds a spreadsheet showing a leaderboard, daily profit and loss, and account growth over time. Every strategy also posts its trades and errors to its own channel in Discord, the chat app, so problems are visible at a glance. One of the fifteen strategies is simply buying and holding a stock index fund, included specifically so every other, more complex strategy has something honest to be compared against. The project documents several real reliability lessons learned the hard way, such as a bug where market data was fetched in the wrong order and looked frozen, and text file formatting issues from Windows computers quietly breaking configuration values. Everything runs on Python, using the Alpaca brokerage API for paper trading, systemd for supervising the services, and Discord webhooks for alerts.