Release notes

• Added support for MongoDB 8.0 API drivers.

• Added support for Planner Version3 that extends performance improvements to aggregation stage operators, along with supporting aggregation pipeline optimizations and distinct commands.

• Added Support for dictionary-based compression through the Zstandard compression algorithm, improving compression ratio by 5x.

• Added new capabilities: Amazon DocumentDB 8.0 supports collation and views.

• Added support for new aggregation stages and operators: $replaceWith, $vectorSearch, $merge, $set, $unset, $bucket, and 3 new aggregation operators $pow, $rand, $dateTrunc.

• Added support for a new version of text index: Text index v2 in Amazon DocumentDB 8.0 introduces additional tokens, enhancing text search capabilities.

• Includes vector search improvements through parallel vector index build reducing index build time by up to 30x.

• Added support for longer collection name (namespace length up to 255 chars).

• Introduced a new query planner (planner version 2.0). For more information, see Query planner v2 in this guide.

• Added new fields in collstats for "MVCCIDStats" and "gcRuntimeStats" that provide information into the overall health of the collection with respect to garbage collection and aging collections.

• Added support of $regexFindAll aggregation operator.

• Index bloat and collection based bloat metrics launched

• Added CloudWatch metrics AvailableMVCCIds and LongestActiveGCRuntime that provide information into the overall health of the cluster with respect to garbage collection and aging collections.

• The serverStatus command now includes an "nvme_cache" field that shows the count of pages written and not written to NVMe cache on NVMe-backed instances.

• Removed limitation on number of fields for $group and $project

• Support for the following String operators: $trim, $rtrim, $ltrim, $regexFindAll, $replaceOne, and $replaceAll

• Support for the following Array operators: $first and $last

• Support for the following Stage operator: $collStats

• Fixed killOp command to handle a special case of bulk inserts.

• Improved network I/O-related memory usage on Amazon DocumentDB instances.

• Fixed count command for $text filter queries.

• Added support for minimum TLS versions tls1.2+ and tls1.3+ to tls cluster parameter.

• Enabled support for characters dollar($) and dot(.) in field names. For functional differences, see Dollar($) and dot(.) in field names.

• $setOnInsert now supports the _id field during inserts resulting from update operations.

• Fixed issue that prevents reclaiming storage space after a major version upgrade to Amazon DocumentDB 5.0.

• Added support for longer index names (up to 255 characters). For more information, see Naming constraints

• Increased the maximum connection limit by two-fold:

  Instance type	Original limit	New limit
  t3.medium	500	1000
  t4g.medium	500	1000
  r5.large	1700	3400
  r5.xlarge	3500	7000
  r5.2xlarge	7100	14200
  r5.4xlarge	14200	28400
  r5.8xlarge	28400	60000
  r5.12xlarge	30000	60000
  r5.16xlarge	30000	60000
  r5.24xlarge	30000	60000
  r6g.large	1700	3400
  r6g.xlarge	3500	7000
  r6g.2xlarge	7100	14200
  r6g.4xlarge	14200	28400
  r6g.8xlarge	28400	60000
  r6g.12xlarge	30000	60000
  r6g.16xlarge	30000	60000

  For more information, see Instance limits.

• Added support DML Audit filtering. You can now setup filter conditions to filter out DML Audit logs based on their specific requirements instead of logging every DML query. For more information, see Filtering DML audit events.

• Added support in document compression for the following:

  • Setting a minimum compression threshold

  • Enabling compression for existing collections (applicable to new documents)

  • Allow default compression setting at the cluster level

  For more information, see Managing collection-level document compression.

• Added support for consuming change streams on reader instances. For more information, see Using change streams on secondary instances.

• Added support for $regexMatch and $regexFind operators.

• Added support to ensure full precision in audit logs when addressing large integers. Audit logs now maintain the exact numeric representation for all numbers, preventing any loss of precision.

• Added support for bypassDocumentValidation and granular error message for $jsonSchema. For more information about bypassDocumentValidation, see bypassDocumentValidation.

• Added support of $expr.

• Added support for Uncorrelated Joins in $lookup.

• Added support to retain validation rules in $out aggregation stage.

• Added support for HNSW vector indexes. For more information, see Vector search for Amazon DocumentDB.

• Added a support for partial indexes. For more information, see Partial index.

• Added a support for GC runtime on a collection within currentOp command.

• Added text index support for native text search on Amazon DocumentDB. For more information, see Performing text search with Amazon DocumentDB.

• Added support for $jsonSchema schema keywords type, allOf, oneOf, anyOf, not, maxItems, minItems, maxProperties, minProperties, pattern, patternProperties, multipleOf, dependencies, and uniqueItems.

  For more information see Using JSON schema validation.

• Added support for arithmetic operators $ceil, $floor, $ln, $log, $log10, $sqrt, and $exp.

  For more information see Arithmetic operators.

• Added support for the conditional expression operator $switch.

• Added support for parallel IVFFLAT vector index builds. Documentation was updated by removing the parallel IVFFLAT vector index builds limitation from the developer guide.

• Added back case-insensitive functionality for invoking db.runCommand("dbstats"). Amazon DocumentDB 5.0customers on engine patch versions prior to 3.0.4960 or 2.0.10221 should apply these latest engine patches.

• Fixed an error when invoking db.coll.stats() on mongo shell version 1.7 and later. Documentation was updated by removing the mongo shell db.coll.stats() troubleshooting tip from the developer guide.

• Added support for $jsonSchema schema keywords maxLength, minLength, maximum, minimum, exclusiveMaximum, exclusiveMinimum, items, and additionalItems.

  Please note that JSON schema validation is supported in instance-based clusters only.

• Added support for $convert aggregation pipeline operator and its shorthand derived operators $toBool, $toInt, $toLong, $toDouble, $toString, $toDecimal, $toObjectId, and $toDate.

• Added support for set expression operators $setDifference, $anyElementTrue, and $allElementTrue.

• Added support for document compression.

• Added support for parallel index builds.

• Added support for index build status.

• Fixed authentication issue with createCollection for Amazon DocumentDB elastic clusters when users don’t have access to system collections.

• Fixed issue where secondary region instances couldn’t use the same primary region instance names.

• r5 and t3.medium instances are now supported in Amazon DocumentDB 5.0.

• engineVersion option default is 5.0.0 in AWS SDK, AWS CLI, and CloudFormation.

• Added support for ignoreunknownindexoptions in the createIndex command.

• Stopped adding internal monitoring queries to auditing logs.

• User defined local variable names now support “_” (underscore) for projection operators such as $let and $filter.

• Introduced Amazon DocumentDB 5.0

  • MongoDB 5.0 compatibility (support for MongoDB 5.0 API drivers)

  • Support for Client-side Field Level Encryption (FLE). You can now encrypt fields at the client-side before writing the data to Amazon DocumentDB cluster. For more information, see Client-side field level encryption

  • New aggregation operators: $dateAdd, $dateSubtract

• Increased storage limit to 128 TiB for all instance-based Amazon DocumentDB clusters and shard-based elastic clusters.

• Amazon DocumentDB 5.0 now supports index scan with the $elemMatch operator in the first nesting level. Index scans are supported when query only have one level of the $elemMatch filter and the nested $elemMatch query does not support index scan.

  Query shape that supports index scan:

  db.foo.find( { "a": {$elemMatch: { "b": "xyz", "c": "abc"} } })

  Query shape that does not support index scan:

  db.foo.find( { "a": {$elemMatch: { "b": {$elemMatch: { "d": "xyz", "e": "abc"} }} } })

• Index bloat and collection based bloat metrics launched

• Added CloudWatch metrics AvailableMVCCIds and LongestActiveGCRuntime that provide information into the overall health of the cluster with respect to garbage collection and aging collections.

• Support for the following String operators: $trim, $rtrim, and $ltrim

• Support for the following Stage operator: $collStats

• Fixed killOp command to handle a special case of bulk inserts.

• Improved network I/O-related memory usage on Amazon DocumentDB instances.

• Added support for minimum TLS versions tls1.2+ and tls1.3+ to tls cluster parameter.

• $setOnInsert now supports the _id field during inserts resulting from update operations.

• Fixed issue that prevents reclaiming storage space after a major version upgrade to Amazon DocumentDB 5.0.

• Added support for longer index names (up to 255 characters). For more information, see Naming constraints

• Increased the maximum connection limit by two-fold:

  Instance type	Original limit	New limit
  t3.medium	500	1000
  t4g.medium	500	1000
  r5.large	1700	3400
  r5.xlarge	3500	7000
  r5.2xlarge	7100	14200
  r5.4xlarge	14200	28400
  r5.8xlarge	28400	60000
  r5.12xlarge	30000	60000
  r5.16xlarge	30000	60000
  r5.24xlarge	30000	60000
  r6g.large	1700	3400
  r6g.xlarge	3500	7000
  r6g.2xlarge	7100	14200
  r6g.4xlarge	14200	28400
  r6g.8xlarge	28400	60000
  r6g.12xlarge	30000	60000
  r6g.16xlarge	30000	60000

  For more information, see Instance limits.

• Added support DML Audit filtering. You can now setup filter conditions to filter out DML Audit logs based on their specific requirements instead of logging every DML query. For more information, see Filtering DML audit events.

• Added support to ensure full precision in audit logs when addressing large integers. Audit logs now maintain the exact numeric representation for all numbers, preventing any loss of precision.

• Added support for bypassDocumentValidation for $jsonSchema. For more information about bypassDocumentValidation, see bypassDocumentValidation.

• Added support of $expr.

• Added support for Uncorrelated Joins in $lookup.

• Added support to retain validation rules in $out aggregation stage.

• Added a support for GC runtime on a collection within currentOp command.

• Added support for $jsonSchema schema keywords type, allOf, oneOf, anyOf, not, maxItems, minItems, maxProperties, minProperties, pattern, patternProperties, multipleOf, dependencies, and uniqueItems.

  For more information see Using JSON schema validation.

• Added support for arithmetic operators $ceil, $floor, $ln, $log, $log10, $sqrt, and $exp.

  For more information see Arithmetic operators.

• Added support for the conditional expression operator $switch.

• Added support for $jsonSchema schema keywords maxLength, minLength, maximum, minimum, exclusiveMaximum, exclusiveMinimum, items, and additionalItems.

  Please note that JSON schema validation is supported in instance-based clusters only.

• Added support for $convert aggregation pipeline operator and its shorthand derived operators $toBool, $toInt, $toLong, $toDouble, $toString, $toDecimal, $toObjectId, and $toDate.

• Added support for set expression operators $setDifference, $anyElementTrue, and $allElementTrue.

• Added support for parallel index builds.

• Stopped adding internal monitoring queries to auditing logs.

• Fixed issue with DML auditing when it is enabled during an ongoing workload.

• Fixed issue with DML auditing when aggregate commands with hint is passed a string value.

• Fixed issue with listCollections command not working when users with readwriteanydatabase role having both authorizedCollections and nameOnly options set to true.

• Fixed issue to properly parse numeric string in a field name.

• Cancel long running cursors when they are impacting garbage collection.

• User defined local variable names now support “_” (underscore) for projection operators such as $let and $filter.

4.0

• Added support for Decimal128 data type. The Decimal128 is a BSON data type supported in all regions where DocumentDB is available.

  For more information, see Data Types.

• Added support for DML query auditing with Amazon CloudWatch Logs. Now Amazon DocumentDB can record Data Manipulation Language (DML) events and Data Definition Language (DDL) events to Amazon CloudWatch Logs.

  For more information, see this blog post.

4.0

• You can now change your own passwoprd with own password with changeOwnPassword privilege.

4.0 (Engine Patch Version 2.0.5170)

• Amazon DocumentDB now offers a free trial. For details, see Amazon DocumentDB free trial page.

• You can now use enhanced features with Geospatial query, including the following APIs:

  • $geoWithin

  • $geoIntersects

• Added support for the following MongoDB operators:

  • $mergeObjects

  • $reduce

  For more information, see the Querying Geospatial data with Amazon DocumentDB.

Amazon DocumentDB 4.0 (Engine Patch Version 2.0.5706)

• Amazon DocumentDB Graviton2 (r6g.large, r6g.2xlarge, r6g.4xlarge, r6g.8xlarge, r6g.12xlarge, r6g.16xlarge and t4g.medium) instances are now supported.

4.0 (Engine Patch Version 2.0.5706)

• Added support for the following MongoDB APIs:

  • $reduce

  • $mergeObjects

  • $geoWithin

  • $geoIntersects

4.0 (Engine Patch Version 2.0.5704)

• Added support for the following MongoDB APIs

  • $literal

  • $map

  • $$ROOT

• Support for GeoSpatial Query capabilities. See this blog post for more details

• Support for access control with user-defined roles. See this blog post for more details

• Amazon DocumentDB JDBC Driver to enable connectivity from BI tools such as Tableau and query tools such as SQL Workbench

4.0 (Engine Patch Version 2.0.5704)

• Bug fix for $natural to sort correctly when an explicit .sort() is present along with $natural

• Bug fix for change stream to work with $redact

• Bug fix for $ifNull to work with empty array

• Bug fix for excessive resource consumption/server crash when a currently logged-in user is deleted or that user’s privilege for an ongoing activity is revoked

• Bug fix in listDatabase and listCollection privilege check

• Bug Fix dedupe logic for multi-key elements

4.0 (Engine Patch Version 2.0.3371)

• r5.8xlarge and r5.16xlarge instances are now supported. Learn more at the blog post Amazon DocumentDB Now Supports r5.8xlarge and r5.16xlarge Instances.

• Global clusters are now supported to provide disaster recovery from region-wide outages and enable low-latency global reads by allowing reads from the nearest Amazon DocumentDB cluster. Note that global clusters are not currently supported in the South America (São Paulo), Europe (Milan), China (Beijing), and China (Ningxia) Regions.

4.0 (Engine Patch Version 2.0.3371)

• renameCollection

• $zip

• $indexOfArray

• $reverseArray

• $natural

• $hint support for update

• Index scan for distinct

4.0 (Engine Patch Version 2.0.3371)

• Reduced memory usage for $in queries

• Fixed a memory leak in multikey indexes

• Fixed the explain plan and profiler output for $out

• Added a timeout for operations from internal monitoring system to improve reliability

• Fixed a defect impacting the query predicates passed to multikey indexes

Amazon DocumentDB 4.0 (Engine Patch Version 2.0.722)

• None

Amazon DocumentDB 4.0 (Engine Patch Version 2.0.722)

• MongoDB 4.0 compatibility

• ACID transactions

• Support for cluster(client.watch() or mongo.watch()) and the database level (db.watch()) change streams

• Ability to start or resume a change streams using startAtOperationTime

• Extend your change stream retention period to 7 days (previously 24 hours)

• AWS DMS target for Amazon DocumentDB 4.0

• CloudWatch metrics: TransactionsOpen, TransactionsOpenMax, TransactionsAborted, TransactionsStarted, and TransactionsCommitted

• New fields for transactions in currentOp, ServerStatus, and profiler.

• Ability to use an index with the $lookup aggregation stage

• find() queries with projections can be served direction from an index (covered query)

• Ability to use hint() with the findAndModify

• Performance optimizations for $addToSet operator

• Improvements to reduce overall index sizes.

• New aggregation operators: $ifNull, $replaceRoot, $setIsSubset, $setInstersection, $setUnion, and $setEquals

• With the ListCollection and ListDatabase commands, you can now optionally use the authorizedCollections and authorizedDatabases parameters to allow users to list the collections and databases that they have permission to access without requiring the listCollections and listDatabase roles, respectively

• Users can also end their own cursors without requiring the KillCursor role

• Comparing numeric types of subdocuments is now consistent with comparing numeric types of first-level documents. The behavior in Amazon DocumentDB 4.0 is now compatible with MongoDB.

Amazon DocumentDB 4.0 (Engine Patch Version 2.0.722)

• $setOnInsert no longer allow updates when using the positional operator $. The behavior in Amazon DocumentDB 4.0 is now compatible with MongoDB.

• Fixed issue with $createCollection and set autoIndexId

• Projection for nested documents

• Changed default setting for working memory to scale with instance memory size

• Garbage collection improvements

• Lookup with empty key in path, behavior difference with mongo

• Fixed dateToString bug in timezone behavior

• Fixed $push (aggregation) to respect sort order

• Fixed bug in $currentOp with aggregate

• Fixed issue with readPreference on secondary

• Fixed issue with validating $createIndex is the same database as the command was issued

• Fixed inconsistent behavior for minKey, maxKey lookup fails

• Fixed issue with $size operator not working with composite array

• Fixed issue with the negation of $in with regex

• Fixed issue with $distinct command run against a view

• Fixed issue with aggregations and find commands sorting missing fields differently

• Fixed $eq to regular expression not checking type

• Fixed $currentDate bug in timestamp ordinal position behavior

• Fixed millisecond granularity for $currentDate

• Fixed issue with DML auditing when it is enabled during an ongoing workload.

• Fixed issue with DML auditing when aggregate commands with hint is passed a string value.

• Fixed issue with listCollections command not working when users with readwriteanydatabase role having both authorizedCollections and nameOnly options set to true.

• Fixed issue to properly parse numeric string in a field name.

• Cancel long running cursors when they are impacting garbage collection.

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.208152)

• Added support for Decimal128 data type. The Decimal128 is a BSON data type supported in all regions where DocumentDB is available.

  For more information, see Data Types.

• Added support for DML query auditing with Amazon CloudWatch Logs. Now Amazon DocumentDB can record Data Manipulation Language (DML) events and Data Definition Language (DDL) events to Amazon CloudWatch Logs.

  For more information, see this blog post.

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.208152)

• You can now change your own passwoprd with own password with changeOwnPassword privilege.

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207684)

• Amazon DocumentDB now offers a free trial. For details, see Amazon DocumentDB free trial page.

• You can now use enhanced features with Geospatial query, including the following APIs:

  • $geoWithin

  • $geoIntersects

• Added support for the following MongoDB operators:

  • $mergeObjects

  • $reduce

  For more information, see the Querying Geospatial data with Amazon DocumentDB.

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207781)

• Added support for the following MongoDB APIs:

  • $reduce

  • $mergeObjects

  • $geoWithin

  • $geoIntersects

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207780)

• Added support for the following MongoDB APIs

  • $literal

  • $map

  • $$ROOT

• Support for GeoSpatial Query capabilities. See this blog post for more details

• Support for access control with user-defined roles. See this blog post for more details

• Amazon DocumentDB JDBC Driver to enable connectivity from BI tools such as Tableau and query tools such as SQL Workbench

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207780)

• Bug fix for $natural to sort correctly when an explicit .sort() is present along with $natural

• Bug fix for change stream to work with $redact

• Bug fix for $ifNull to work with empty array

• Bug fix for excessive resource consumption/server crash when a currently logged-in user is deleted or that user’s privilege for an ongoing activity is revoked

• Bug fix in listDatabase and listCollection privilege check

• Bug Fix dedupe logic for multi-key elements

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207117)

• r5.8xlarge and r5.16xlarge instances are now supported. Learn more at the blog post Amazon DocumentDB Now Supports r5.8xlarge and r5.16xlarge Instances.

• Global clusters are now supported to provide disaster recovery from region-wide outages and enable low-latency global reads by allowing reads from the nearest Amazon DocumentDB cluster. Note that global clusters are not currently supported in the South America (São Paulo), Europe (Milan), China (Beijing), and China (Ningxia) Regions.

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207117)

• renameCollection

• $zip

• $indexOfArray

• $reverseArray

• $natural

• $hint support for update

• Index scan for distinct

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.207117)

• Reduced memory usage for $in queries

• Fixed a memory leak in multikey indexes

• Fixed the explain plan and profiler output for $out

• Added a timeout for operations from internal monitoring system to improve reliability

• Fixed a defect impacting the query predicates passed to multikey indexes

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• Ability to use an index with the $lookup aggregation stage

• find() queries with projections can be served direction from an index (covered query)

• Ability to use hint() with the findAndModify

• Performance optimizations for $addToSet operator

• Improvements to reduce overall index sizes

• New aggregation operators: $ifNull, $replaceRoot, $setIsSubset, $setInstersection, $setUnion, and $setEquals

• Users can also end their own cursors without requiring the KillCursor role

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• None

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• None

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• Added the ability to open a change stream cursor at the cluster level (client.watch() or mongo.watch()) and the database (db.watch())

• Ability to increase the change stream retention period to 7 days(previously 24 hours)

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• Various general case performance improvements

• A targeted security improvement

• Fixed an issue with skip sort on second field of a compound index

• Enable regular index for equality on single field of a multi-key index (not compound)

• Fixed authentication race condition

• Fixed issue that caused an infrequent garbage collection crash

• RBAC security improvement

• Added databaseConnectionsMax metric

• Performance improvements for certain workloads on r5.24xlarge instances

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• $out aggregation stage

• Increased the maximum number of connections and cursor per instance by as much as 10x

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• None

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• Cross Region Snapshot Copy

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• None

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• T3 medium instances

Amazon DocumentDB 3.6 (Engine Patch Version 1.0.206295)

• Idle memory reclamation for t3 instances

• Authentication improvements

• Improved SASL authentication performance

• Fixed currentOp issue when exceeding maximum possible ops

• Fixed killOps issue for bulk update and delete

• Improvements to $sample performance with $match

• Fixed support for $$ in cond case in redact stage

• Fixed various recurring crash root causes

• Improvements to TTL sweeping to reduce IOs and latency

• Optimized memory utilization for $unwind

• Fixed collection stats race condition with drop index

• Fixed race condition during concurrent index build

• Fixed infrequent crash in hash_search in index