Hi... sorry for these simple questions:
What's your DRS automation level and migration thresholds?
DRS is set in the middle (i.e. Apply Priority 1, 2 and 3 recommendations).
In addition to adding more RAM to ESXi Hosts, is there any other thing we can do in the mean time ? Like: Turn off some VMs / set up Resource Pools ?
... and what about automation? Is this in fully automated? Try to increase threshold... then check memory/processor availability of the other hosts.
To add more available RAM you should: poweroff or decrease VM memory and/or set a limit in resource pool. Then scale-in or out the cluster (adding physical RAM or adding another host).
Another thing you could to do is set more shares in a resource pool which contains critical VM.