I'd like to know how Stack Exchange is using ElasticSearch in practice (if it's not a secret):
- How did you organize the indexes? (based on types, on time, something else?)
- How many shards and replicas are used per index?
- How many nodes you have?
- How many documents are in one index?