have a configurable timeout between nodes restart in rolling restart playbooks

[tarzanek]

Regarding cache warmup - it might make sense to have a sleep value between node restarts to allow node to start and warm up (the cache)

this can result into less impact for the clients if cluster is highly utilized

[tarzanek]

https://github.com/scylladb/scylla-ansible-roles/blob/master/example-playbooks/rolling_ops/rolling_restart.yml

would get a configurable delay or sleep similar other rolling operations can get this, too