I'm a little confused about the pricing model for Elastic Cloud. Specifically, I don't understand why I would ever subscribe to platinum over enterprise. It appears with the inclusion of searchable snapshots in enterprise, it is more cost effective to cut down to a single cold/frozen node with snapshot redundancy as opposed to platinum with two zones for cold and frozen data.
With this, I am also a little confused about how cold and frozen data work with searchable snapshots. Do you even need a zone to represent that data or can it all be stored within the snapshot? The way the documentation and articles on the subject read, it sounds like cold data needs a node, but frozen data can be completely stored within a searchable snapshot. Am I misunderstanding this?
There is an uplift for Enterprise Subscription and it may not make sense for all use cases... like Search use cases that have a low volume of data but a very high search rates that needs low latency.
And Yes enterprise for Telemetry or Observability Data can be very cost effective that is why searchable snapshots are included in that tier.
There is quiet a lot of detail here but I will try to summarize
Cold : Has a the Primary Replica of the Data on Local Disk and it is backed by a searchable snapshot if any shard is lost or corrupted in can be quickly restored. There is no replica data. The Cold Node in Elastic Cloud holds about 10TB of data backed by a searchable snapshot
Frozen : Frozen does require a frozen node. That nodes uses a smart cache to hold the most recently access data and also is the entry point for the Snapshots that stored in s3 compatible storage. We currently size a Frozen Node to access ~100TB of s3 Snapshots.
Here is an article as well
tl;dr Yeah Frozen is pretty awesome... test it out!
That's all super fascinating. So can someone with a platinum subscription take full advantage of the frozen tier? It seems pretty much limited to enterprise in that it requires searchable snapshots. Or am I misunderstanding?
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.