the chosen database for our current project is Elasticsearch as it is handy enough for JSON documents and our project involves only JSON documents (data-wise).
I'm a bit concerned because no one in the company is fluent in elasticsearch. I have questions regarding 2 things:
- as I understand, we will have only a VM to deploy all of our modules (.jar mostly) and elasticsearch. How much RAM and CPU should we give to ES in a context where at most 60.000 requests could be sent for create/update and between [0-60.000] requests could be sent for search/get (this 60.000 number is conservative, I would expect something like a peak load of 10.000 users maximum, thus resulting in approximately 10.000 get/search requests)
- I know snapshots can be scheduled every half hour or so to secure the data (that frequency seems ok to me, relative to the value of the data).
- How do I know how to size my elasticsearch database ? What are the main parameters to consider ? For example when I test on my machine I see health is YELLOW and for example I don't know how to go from there to GREEN.
Sorry this post might be messy but I want to make sure elasticsearch is a suitable solution as a primary database, otherwise we need to look somewhere else.