Hello All : I have some basic questions that I hope will help me get a better understanding of migrating our EC2 based stack to ECS. Any Input will help feed my understanding! Please note, I am NOT using Fargate ECS. J
This is our current configuration (Just trying to do migrate elasticsearch for now):
3 Coordinating Nodes 3 EC2 Instances each Instance has 64 GB memory 7 Master/Data Nodes 7 EC2 Instances each Instance has 64 GB memory
Limiting myself on how to migrate the 7 Master Data Nodes above into ECS : If I were to attempt to map the master data Nodes to ECS and achieve the same level of SLAs and Response/Up times maintained after migrating to ECS - what should my ECS Cluster look like?
- Do I spin up 7 EC2 Instances in my ECS cluster to maintain the same SLA? (ie choose DAEMON when creating the Service?)
- When creating the container definition, what is the hard memory limit I should specify? (Currently we specify 30 GB since we have 64GB as suggested by elastic for optimal performance. of course, we have 1 elastic service running on 1 EC2 Instance. The confusion is how to map this notion to ECS?)