Submitted by Jorge Marin
If you still struggle to understand AWS ECS memory and CPU reservation, or you fear running out of memory, or you thought you had more memory but your tasks still get killed. If you are always 100% CPU. Let's unravel this together and set a list of best practices for maintaining your ECS clusters
In this talk we will have a high level and low level view at AWS ECS, memory and CPU reservation/utilization and how different runtimes (.NETCore, NodeJS) deal with these memory and CPU restrictions. How to scale the cluster properly, how to reschedule the tasks across the new cluster instances. How to properly monitor the memory and the CPU without using misleading metrics or complicated calculations.