Determine Cluster Size Needed

Hello,

I need to determine the pricing to host our future Elasticsearch server on Elastic Cloud. I cannot test this during the 14 days trial period of Elastic Cloud because my boss needs this information as soon as possible and I am totally new to Elasticsearch so it would take me time to have a system running.

I am trying to know how big of a cluster size I would need (to then determine the price).

I hope the following information will be relevant to help me getting close to an estimate?

-I have a Drupal (CMS) website and I would like to use Elasticsearch for a standard website search page. It would mainly be a keyword search and a small faceted search for blogs/reports/news.
-We have around 33,500 queries per month (400,000 per year) made by visitors on our site.
-We can run indexing process once a day.
-We have 50,000 pages (blog posts, reports, and news posts) on our website that should be indexed and around 5000 documents (mainly images and PDFs). I do not know how much these would weight once in JSON format. I read that on average, a record is 1.5kB. I sometimes have long report texts so maybe this number can be increased a bit.
-Pretty much all the fields of my records should be indexed. Long text of documents should be indexed.
-For the latency requirement when querying: the user can wait up to 3 seconds before seeing the search results.
-For the data retention period: I am not sure to understand what that is and how it works. I would need the whole site to be searchable. Is it related to caching?
-For the limitations for hardware storage: I am not sure I understand this part either but all documents are stored on the website host server.

Based on these information, how big do you think would my cluster be? I thought it would either be 4GB/96GB or 8GB/192GB?

Thank you for your help!
Emeline

I'd say it would be larger due to your response SLAs (the 3 sec and the PDFs that you want to index.

It's hard to say how large though. It may make sense to start on the 8GB cluster and then increase things as needed. Make sure you also use HA.

Hello,

Thank you for your input! This is helpful.

What do you mean by using HA (for the configuration of the server)?

Thank you!
Emeline

2 datacenters.

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.