Searching in several documents, I could not find figures that helps sizing your Elasticsearch Cluster for Production.
Here are some findings after indexing Production data, that will help understand better the capacity you need in term of storage:
General Ledger | Storage required for Base Language = 1.1 GB = 670'000 documents in Elasticsearch
- JRNL_HEADER - Journal Header Data - 1'200'00 lines
- JRNL_LN - Journal Line Data - 100'000'000 lines
Account Payabale | Storage required for Base Language = 0.5 GB = 330'000 documents in Elasticsearch
- VOUCHER - AP Voucher Header Table - 3'000'000
- VOUCHER_LINE - Voucher Line - 6'500'000
Payments | Storage required for Base Language = 1.2 GB = 480'000 documents in Elasticsearch
- PAYMENT_TBL - AP Disbursements - 3'000'000 lines
Vendors | Storage required for Base Language = 1.0 GB = 200'000 documents in Elasticsearch
- Vendor - Vendor Header Table - 280'000 lines
So anticipate to get a cluster we enough CPU and RAM, but standard server in any Cloud would be enough in term of storage.