rust-prometheus VS Sphinx

I used the prometheus library because it was the most popular and easiest to find. Some unusual and unexpected aspects were that, to delete metrics, all labels must be specified at once, and to record metrics, I always needed the full list of labels that were used during creation. However, these issues were manageable.

[dependencies] prometheus = { git = "https://github.com/tikv/rust-prometheus", rev = "ac86a264223c8d918a43e739ca3c48bb4aaedb90", features = ["process"] }

Metrics exposition is implemented with prometheus crate.

The new Elasticsearch exporter is written in the Rust programming language and is open-sourced on GitHub: github.com/vinted/elasticsearch-exporter-rs. The exporter uses asynchronous Tokio runtime, Rust Prometheus instrumentation library and the official Elasticsearch client library. Metrics collection is decoupled from the serving /metrics endpoint. In addition, Elasticsearch time-based metrics in milliseconds are converted into seconds to comply with Prometheus best practices (metrics ending in “millis” are replaced by “seconds”, “_bytes” and “_seconds” and postfixes are added where appropriate).

Rust has a popular binding: https://github.com/tikv/rust-prometheus

Sphinx - https://sphinxsearch.com

Sphinx is a search engine that can be integrated into a website to provide advanced search functionality such as full-text, Boolean, and faceted search. It is a powerful open-source search engine that can handle large amounts of data and quickly return results.

Have been using Sphinx. It does some processing around suffixes, tenses, and so on, and looks at word proximity (BM25), but is definitely limited.

Lucene is the thing you think you need. Elastic Search is a nice wrapper for it. But these are Java, so maybe you want Sphinx Search (C++) or MeiliSearch (Rust).

Using a natural language search will almost certainly be a better solution and PHP may not be the best tool for this task. Figure out how you are going to get the text out of the PDF and where you are going to put it. Look at things like sphinx and full text search in boolean mode for doing the keyword matching.

In practice though you don't do any of this, you get a library to do it for you. I've used Sphinx Search in the past for some fairly hefty (In the order of terabytes), and there's a good book covering how to get it all set up and started.

Five years ago Manticore began as a fork of an open source version of the once popular search engine Sphinx Search. We had two bags of grass, seventy-five pellets of mescaline, three C++ developers, a support engineer, a power user of Sphinx Search / backend team lead, an experienced manager, a mother of five helping us part-time, and a ton of bugs, crashes, and technical debts. So we got a shovel and other digging tools and started working to get it up to the search engine industry standards. Not that Sphinx was impossible to use, but many things were missing, and existing features weren’t quite stable or mature. And we had pushed it about as far as we could. So after 5 years and hundreds of new users, we’re ready to say that Manticore Search can be used as an alternative to Elasticsearch for both full-text search and (now) data analytics too.

Finally if you really want a full fledged solution the most popular are good old Sphinx http://sphinxsearch.com or Apache Solr https://solr.apache.org but as these take a bit of time to setup I'd make sure that's truly needed. Chances are unless your collection is the size of the Library of Congress you are probably fine with ripgrep on your own indexes as text or some R package or sqlite FTS5 extension https://www.sqlite.org/fts5.html

An old fact from 2014 December 12th was significant for Vinted: the company switched from Sphinx search engine to Elasticsearch 1.4.1. At the time of writing this post, we use Elasticsearch 7.15. Without a doubt, a lot has happened in between. This chapter will focus on Elasticsearch metrics, we will share our accumulated experiences from four generations of collecting metrics.