[Queue Sinks] added Kafka and RabbitMQ sink. More here
[Data Archival] added 'data archival' feature. More here
Upgrading from previous versions
4.x and 5.x licenses will not work with 6.x products and need to be converted via dedicated tool available here. After conversion license will continue working with previous versions of the product, but can be also used with 6.x ones.
please refer to our Server migration guide before proceeding with Server update and check our list of Server breaking changes available here and Client API breaking changes available here
Server
[Backups] switched FTP backup implementation to use 'FluentFTP'
[Configuration] changed default 'Indexing.OrderByTicksAutomaticallyWhenDatesAreInvolved' value to 'true'
[ETL] OLAP ETL uses latest Parquet.Net package
[ETL] removed load error tolerance
[Graph API] removed support
[Indexes] new auto indexes will detect DateOnly and TimeOnly automatically
[Indexes] added the ability to 'test index'. More here
[JavaScript] updated Jint to newest version
[Monitoring] added OIDs to track certificate expiration and usage
[Querying] when two boolean queries are merged in Lucene, boosting should be taken into account properly to avoid merging queries with different boost value
[Voron] performance improvements
Client API
[Compare Exchange] added support for creating an array as a value in 'PutCompareExchangeValueOperation'
[Compare Exchange] compare exchange includes should not override already tracked compare exchange values in session to match behavior of regular entities
[Conventions] HttpVersion was switched to 2.0
[Conventions] removed 'UseCompression' and introduced 'UseHttpCompression' and 'UseHttpDecompression'
[Conventions] introduced 'DisposeCertificate' with default value set to 'true' to help users mitigate the X509Certificate2 leak. More info here
[Database] introduced 'DatabaseRecordBuilder' for more fluent database record creation
[Facets] removed FacetOptions from RangeFacets
[Graph API] removed support
[Patching] JSON Patch will use conventions when serializing operations
[Session] private fields no longer will be picked when projecting from type
[Session] taking into account 'PropertyNameConverter' when querying and determining field names
[Session] when a document has an embedded object with 'Id' property we will detect that this is not root object to avoid generating 'id(doc)' method there for projection
[Session] no tracking session will throw if any includes are used
removed obsoletes and marked a lot of types as sealed and internal
changed a lot of count and paging related properties and fields from int32 to int64
Studio
[Dashboard] removed Server dashboard
[Design] refreshed L&F
Test Driver
added 'PreConfigureDatabase' method
Other
[Containers] docker will use non-root user. More info and migration guide here
Skip to Navigation