Strategy for Laravel scout with million+ records and multiple interchangeable drivers (i.e. TNT & Algolia)

Question

I have used Algolia a bit and it is an awesome service. I have also used the TNT-search driver for scout and it is also pretty good, but not really a touch on the features, speed and ease that you get with Algolia.

Unfortunately, Algolia gets very expensive when dealing with a lot of records - for example one of our apps has over 10 million searchable rows, which would be thousands of $$ monthly!

Has anyone had any success in using both together? i.e. I would like use Algolia for recent records or categories where I need more advanced search capabilities (100k of records) and then use TNT search for the remainder.

EDIT:

Elasticsearch was the answer. It is a little harder to setup but has such great flexibility. I would highly recommend https://github.com/matchish/laravel-scout-elasticsearch to connect it up with Laravel Scout seamlessly.

Answer 1

Well ,AWS Elasticsearch very good and cheap .. you might use it and it's very easy to use and configure with laravel AWS Elasticsearch pricing You may start with r3.large.elasticsearch it will costs you about 180$ per month and if you want to more r3.xlarge.elasticsearch will be amazing too and it will serve you need.

to configure AWS Elasticsearch with laravel you may read this artical How-to integrate-your-Laravel-app-with-Elasticsearch

Answer 2

Elasticsearch is already suggested in another answer as a cost effective alternative. But if you are looking for something similar to Algolia but open source, check out Typesense. It's must easier to set-up and manage and also features like typo correction etc. just works out of the box.

Answer 3

This answer might be late but if there is anyone who still has this problem, Laravel scout has a searchableUsing() method that determines the search engine to use, you can override that to configure different search drivers for different models. This blogpost is a detailed step by step on how to do this