I have a cluster of 100 shards on 100 nodes, with solr 7.5, running in AWS.The
use case is read-dominant, with ingestion performed about once per week. There
are about 84 billion documents in the cluster. It is unused on weekends and
only used during normal business hours M-F.What I do now is after each round of
ingestion, create a new set of AMIs, then terminate each instance.The next
morning, the cluster is restarted by creating a new set of spot requests, using
the most recent AMIs. At the end of the day, the cluster is turned off by
terminating the instances (if no data was changed), or by creating a new set of
AMIs and then terminating the instances.Is there a better way to do this? I'm
not facing any real problems with this setup, but I want to make sure I'm not
missing something obvious.Thanks,Chris
