We containerized our applications, but, fundamentally, nothing about the application really changed. The best way to monitor your application is through the creation of metrics and logs, as we did in the earlier part of this chapter. That said, if you experience some issues with your containers, you may want to search for unhealthy events in the ECS clusters menu:

Those issues are often caused by misconfiguration of the tasks, when you don't provide enough CPU and memory reservation, or by bugs or bad user behavior. For those issues, it is often hard to diagnose them by solely looking at metrics. To help with the diagnosis of these issues, we can turn on logging on our load balancers.