Java: Spring Boot and Elasticsearch?

Elasticsearch — a search engine for full-text search and analytics. Spring Boot connects via Spring Data Elasticsearch.

Simple analogy: Elasticsearch is like Google for your data. You ask "find all documents about Docker" — it finds them instantly, even similar ones. Unlike a database LIKE %keyword% query (which scans every row), Elasticsearch uses an inverted index — like a book's index at the back. If you search for "Docker", it jumps directly to all pages mentioning it, without reading every page.

Why it matters: Relational databases are terrible at full-text search — LIKE %keyword% cannot rank results by relevance, cannot handle synonyms, and slows down as data grows. Elasticsearch was purpose-built for this: it tokenizes text, handles fuzzy matching, supports auto-complete, and returns results ranked by relevance in milliseconds.

Configuration:

1spring:
2  elasticsearch:
3    uris: http://localhost:9200
4    username: elastic
5    password: ${ES_PASSWORD}
6    connection-timeout: 5s
7    socket-timeout: 30s

Document:

1@Document(indexName = "articles")
2public class Article {
3    @Id
4    private String id;
5
6    @Field(type = FieldType.Text, analyzer = "english")
7    private String title;
8
9    @Field(type = FieldType.Text, analyzer = "english")
10    private String content;
11
12    @Field(type = FieldType.Keyword)
13    private String category;
14
15    @Field(type = FieldType.Date)
16    private LocalDateTime publishedAt;
17
18    @Field(type = FieldType.Double)
19    private Double score;
20}

Repository + Custom Query:

1public interface ArticleRepository extends ElasticsearchRepository<Article, String> {
2    List<Article> findByTitleContaining(String text);
3    List<Article> findByCategoryAndPublishedAtBetween(
4        String category, LocalDateTime from, LocalDateTime to);
5
6    @Query("{"bool": {"must": [{"match": {"title": "?0"}}, {"match": {"content": "?0"}}]}}")
7    List<Article> fullTextSearch(String query);
8    
9    @Query("{ "bool": { "must": [ { "match": { "content": { "query": "?0", "operator": "and" } } } ] } }")
10    List<Article> searchWithAndOperator(String query);
11}

ElasticsearchRepository automatically creates the index and maps objects.

Production considerations:

• Index design — design indices around query patterns, not data structure. Denormalize aggressively.
• Shard count — set at index creation time and cannot be changed later. Start with 3-5 shards for most use cases.
• Bulk indexing — use ElasticsearchOperations.bulkSave() for bulk inserts instead of saving one document at a time.
• Not a primary database — Elasticsearch is eventually consistent and not ACID-compliant. Use PostgreSQL for transactional data and ES for search.

When to use:

• Full-text search (product catalog, documents, articles).
• Logs and metrics (ELK Stack: Elasticsearch + Logstash + Kibana).
• Analytics (aggregations: average order value, popular categories).
• Auto-complete and suggestion features.

Do not use for transactional data — it is not a replacement for PostgreSQL.