Running Elasticsearch would require compute, which we don't have. However, how often do you run analysis on the data? Would it make sense to move the data to B2 and put back on s3 when it's needed?
Not sure what you mean that running Elasticsearch requires compute. I'm talking about the snapshot/restore functionality in Elasticsearch. We use it mostly for backups. As far as I can tell, it should be quite possible to implement a B2 plugin for that (Azure, GCS, HDFS, and S3 are available currently).
Got it. Let me look into that. Agree that if you're just taking a snapshot from Elasticsearch and putting into B2, that should be an implementalbe plug-in.