Last checked and updated on 05 Jun 2026
What Is New in MongoDB 7.3
MongoDB 7.3 delivers a set of targeted enhancements focused on querying, security, and operational efficiency. This release builds on the foundation of previous versions with refinements that developers and DBAs will appreciate in production environments.
| Category | Key Changes |
|---|---|
| Querying & Analytics | New $median and $percentile accumulator operators, $search aggregation stage support |
| Security | Queryable Encryption protocol enhancements, new key management integrations |
| Operability | Sharded cluster balancing improvements, new collation support for time series collections |
| Time Series | Granularity and bucket sizing refinements for improved storage efficiency |
What new aggregation operators were added?
MongoDB 7.3 introduces two powerful statistical operators for data analysis. The new $median and $percentile accumulators provide native support for calculating median values and arbitrary percentiles within aggregation pipelines.
These operators work directly on your data without needing complex manual calculations. You can now compute these statistical measures efficiently as part of your aggregation workflow, which is particularly valuable for analytical applications and reporting.
How does Queryable Encryption improve?
The Queryable Encryption protocol gets significant enhancements in version 7.3. These improvements focus on performance optimizations and expanded functionality for encrypted queries while maintaining strict security guarantees.
You'll notice better query performance on encrypted fields and more flexible query patterns. The update also adds support for additional key management services, giving you more options for how you handle encryption keys in your infrastructure.
What sharding improvements should I know about?
Sharded cluster balancing receives attention in this release with smarter data distribution logic. The balancer now makes more informed decisions about when and how to move chunks between shards, reducing unnecessary data movement.
This means less overhead during balancing operations and more stable performance during data redistribution. The improvements are particularly noticeable in clusters with uneven data distribution patterns or rapidly changing workloads.
Are there time series collection enhancements?
Time series collections gain collation support in MongoDB 7.3, allowing for locale-specific sorting and comparison of string data within your time series documents. This is useful for international applications that need language-aware ordering.
The release also includes refinements to automatic bucket sizing and granularity handling. These changes help optimize storage efficiency and query performance for time series data, especially in scenarios with irregular data ingestion patterns.
Does Atlas Search get any updates?
Yes, the $search aggregation stage receives additional capabilities in MongoDB 7.3. The integration between aggregation pipelines and full-text search becomes more seamless, allowing for complex data processing that combines search with other aggregation operations.
This tight integration means you can build more sophisticated search-driven applications without needing to jump between different query systems. The enhanced $search stage supports more search features directly within aggregation pipelines.
FAQ
Can I use $median and $percentile with grouped data?
Yes, both new operators work within $group stages and support computing statistics across grouped documents, making them perfect for analytical queries against categorized data.
Do the Queryable Encryption changes break existing encrypted collections?
No, the enhancements maintain backward compatibility. Existing encrypted collections continue to work without modification while benefiting from the performance improvements.
How do the sharding improvements affect balancer behavior?
The balancer becomes more conservative about initiating chunk migrations, reducing unnecessary data movement while still maintaining balanced cluster distribution during data growth.
What collations are supported for time series collections?
Time series collections now support the same collation options as regular collections, including locale-specific sorting rules for string fields within time series documents.
Can I combine $search with the new statistical operators?
Absolutely. Since $search is an aggregation stage, you can pipe search results directly into stages that use $median or $percentile for sophisticated search analytics pipelines.