My team is using Elastic Cloud (Azure cloud provider) and we recently noticed that the hardware profiles have changed when creating a new deployment. Previously, we had chosen the I/O optimized profile which is no longer present. When looking at the new profiles, one confusing phrase that it uses is "ingestion use cases with 7-10 days of data available for fast access". Can anyone explain to what "available for fast access" is referring? Is that just referring to the RAM and cached results?
In our use case, all of our data is Hot data, not log/time-series data. It's basically record tracking data where any record can be updated at any time. Peak ingestion can be up to millions of records per hour (including updates and new records), peak searches could be in the 30-50 requests per second range. Data is likely in the range of multiple terabytes. We're also making significant use of aggregations across this data to calculate various metrics.
In our internal testing, CPU has been the the typical bottleneck for ingestion/search causing us to scale up and the new hardware profiles (General and CPU optimized) look to provide more CPU resources which looks attractive, but I would just like to understand the intent and meaning behind that "available for fast access" wording before we decide on a new profile.