Redash on ECS - EC2 Instances with AutoScaling

[anshulgarg-07]

Hey guys, up until now we had deployed Redash on a single EC2 instance configured with an RDS and Elasticache (Redis) for metadata persistence. Now, because of hopeful cost savings and more optimised resource distribution, we want to move to AWS ECS (EC2 instances) with auto scaling (server & workers) for production usecase. I will have different services for server, workers and scheduler.

I want to know all possible suggestions & tips you can give me for it. One issue I am anticipating is the handling of running queries if the worker task handling them is descaled. All kind of help regarding this would be highly appreciated.