Overall Satisfaction with Elasticsearch
We use Elasticsearch as a search provider for our ecommerce software. Our search, category and navigation pages are rendered from ElasticSearch.
- More relevant search results. There are lot of in build algorithms that are part of Elasticsearch. Using these algorithms improved search results.
- Decrease in the page load time since read operation is very fast.
- Easy to implement when compared to other software.
- Installation and configuration of Elasticsearch on windows server is not straight forward.
- Completion suggester algorithm in Elasticsearch (v 2.0) saves information in memory. So any deletes/updates are not reflected immediately unless a flush command is executed. Execution of flush command is not advised by Elasticsearch team.
- Elasticsearch Nest API code is not updated to match with Elasticsearch release version. So we have to write our own implementation.
- More relevant search results are displayed to user. So chances of purchasing a product is very likely.
- Managing an elasticsearch server is very easy. No maintenance is required.
- Documentation is very bad. More time is spent on research than implementation.
Even though Lucene is very powerful it is not easy to implement Lucene as a search provider. Lucene is the core of Elasticsearch and they made implementation very easy.