I'm currently developing a platform where users can run Python trading bots, with each strategy executed in its own Docker container. Given that I have 10 users each running 3 strategies, that's a total of 30 containers active at the same time. However, I'm facing some challenges. When a user clicks to stop all strategies, the system becomes sluggish as it tries to shut down all the containers. Additionally, I'm fetching balance updates and other info every 30 seconds, which causes the web interface to run slowly. I'm trying to figure out the best way to scale this system for over 500 users. Should I completely rethink my architecture? I'm currently using an m5.xlarge EC2 instance. Any insights from those with experience building similar systems would be much appreciated!
1 Answer
When you mention the system lags, what exactly do you mean? Systems often require horizontal scalability, but that can lead to higher latency for certain operations like API requests.
I have buttons in the UI to start and stop containers, and when I hit the close button, it tries to remove all containers. During that time, if another request comes in, it sometimes causes delays or instability.