Recommendations for scaling up/down and disk size changes

[vaughnd]

Hi, I'm having trouble with scaling. I can add N nodes fine, and drop 1 at a time and wait for shards to allocate. But if I try to increase disk size and add N nodes then the new nodes don't join the cluster. Is there a way to get the ASG to have some intelligence around scaling elasticsearch safely, mainly when changing disk size? Or is it best to just create a new cluster for disk size changes?

Some guidelines for scaling safely will be much appreciated. Or pointers to docs and code that explain this. Thanks.