Atlas Search is MongoDB Atlas’s embedded full-text search capability, built on Apache Lucene. It adds fast, relevant search, custom scoring, and facets without requiring a separate search system. section 1.3.2 section 11.3
Use Atlas Search when search integration, synchronization, and operational overhead matter:
- Access database and full-text search through one MongoDB Query API and unified driver. section 11.3
- Rely on Lucene indexes that synchronize with MongoDB through the mongot process, which monitors change streams and applies updates. section 11.8
- Choose dynamic mappings to index supported fields automatically, or static mappings to select fields and ignore mismatched types. section 11.8
- Place search workloads on dedicated Atlas Search Nodes for workload isolation on eligible M10-or-higher clusters. section 11.8
Index and deployment choices
An Atlas Search index can cover one field or several fields, especially fields used for sorting or filtering. Dynamic mappings simplify coverage but may increase resource use; static mappings provide selective indexing. M0 clusters allow up to three indexes, while M10-and-higher clusters have no index-count restriction. section 11.8
Figure 11.2 High-level overview of Atlas Search. The application uses MongoDB Atlas with a unified driver and query API to access both database and search functionality. Within the Atlas environment, data storage and search are integrated, eliminating the need for separate systems or synchronization. This unified approach simplifies the architecture and reduces operational overhead. (Image © MongoDB 2024 CC BY-NC-SA 3.0)
Trace how an application uses MongoDB Atlas, a unified driver, and one Query API to reach integrated database and search functionality. section 11.3
