What to choose for multilingual search solr, sphinx, elastic or what else?

Question

Sphinx

Support for Russian, English and Asian languages ​​​​(Chinese, Korean, Japanese) is a prerequisite. We decided to communicate with the service using the REST API.
The data will fall into the index at some intervals from several users.
I have experience using Sphinx and Solr .
The first one has excellent speed and quality of searching in Russian using lemmas, but "difficulties" with rebuilding the index. Not to say that they are insurmountable, but it is much better that all operations are hidden from the end user of the system without any nightly index rebuilds and other things (I know about rt and delta indexes). Didn't use sphinx to search across multiple languages.
The second rest api out of the box (like elastic), work with Russian and other languages ​​according to the stemma. The speed of search and indexing is lower than that of Sphinx, but you do not need to rebuild, merge the index every time, for me this is a huge plus. Haven't used Solr to search across multiple languages.
C elasticSearch didn't work.
Tell me where to stop?

Answer 1

Here you say that you worked with Sphinx, let's start from this. According to your post, I concluded that the only thing that stops you in Sphinx is the imaginary need to stop the service and rebuild the indexes. I hasten to please you - this is not necessary. I advise you to carefully study the official documentation (preferably from cover to cover, it is not as big as it might seem) and get acquainted with the wonderful world of delta-index and rt-index. They will allow you to forget stopping the service to rebuild indexes like a bad dream. Another thing is that not everything is in order with them, which is also described in detail in the documentation.