Overview of 8.0 New Features

Last updated: 2025-11-27 17:55:56

Overview of 8.0 New Features

Last updated: 2025-11-27 17:55:56

For a complete list of changes and details of new features in MongoDB 8.0, see the official MongoDB 8.0 release notes.
Category
Feature Name
Description
Use Case/Impact
Performance optimization
Enhanced TCMalloc
Caching mechanism: changed from per-thread caching to per-CPU caching to reduce memory fragmentation and improve adaptability under high load.
Memory release: a background thread that attempts to release memory to the operating system per second is added. The tcmallocReleaseRate parameter allows precise control of the release rate (in the unit of Bytes/s), with a default value of 10 MB/s.
More efficient memory usage is achieved, and it is especially suitable for scenarios with high memory pressure and significant load fluctuations.
﻿
Majority write optimization
When write concern is set to "majority", acknowledgment is returned once the oplog is written to a majority of nodes, without waiting for the data application.
This significantly improves throughput and reduces latency for majority writes.
﻿
oplog processing optimization
MongoDB 8.0 introduces an oplog buffer on secondary nodes, decoupling the data synchronization process into two independent stages: the Writer thread continuously pulls and buffers oplogs, while the Applier thread asynchronously applies data from the buffer. This allows parallel execution of the two threads, eliminating resource waste caused by serial waiting.
This improves the data replication throughput of secondary nodes and reduces replication latency between primary and secondary nodes.
﻿
Batch insertion optimization
Non-transactional batch insertions are merged into a single oplog entry, with all documents sharing the same clusterTime in the Change Stream.
This improves batch insertion performance and reduces replication latency caused by secondary nodes replaying a large volume of oplog entries.
Shard enhancement
reshard optimization
forceRedistribution: allows for resharding using the same shard key to quickly redistribute data to new shards.
commitReshardCollection: enables proactive disabling of write operations during the Ops window to accelerate the completion of data shard migration.
This provides faster and more flexible data redistribution, which facilitates load balancing and capacity planning.
﻿
Shard management commands
unshardCollection: unshards a collection by merging data into a specified shard.
moveCollection: moves an unsharded collection to any shard (excluding time series and encrypted collections).
shardAndDistributeCollection: redistributes data immediately after sharding.
This provides more granular and flexible collection-level data layout management capabilities.
﻿
Shard operation optimization
Hashed sharding: By default, 1 chunk is created per shard (previously 2 chunks per shard).
findAndModify/deleteOne: Partial shard keys can be used as query conditions.
updateOne with upsert: For sharded collections, the query conditions of the upsert operation do not need to include all the shard keys.
directShardOperations role: Commands are allowed to be run directly on shards for shard operations.
This simplifies Ops and queries for sharded clusters, enhancing usability.
﻿
 serverStatus output enhancement
Starting from MongoDB 8.0, the serverStatus output includes the following new shardingStatistics fields:
shardingStatistics.countTransitionToDedicatedConfigServerStarted
shardingStatistics.countTransitionToDedicatedConfigServerCompleted
shardingStatistics.countTransitionFromDedicatedConfigServerCompleted
shardingStatistics.configServerInShardCache
This provides a more comprehensive view of the running status and internal statistical metrics of sharded clusters.
Query and aggregation
Query statistics and optimization
$queryStats: returns detailed statistics for recorded queries.
Query Shape: introduces a new query shape, upgrading query pattern recognition from a passive cache management mechanism to an independently configurable and manageable resource.
workingMillis: displays actual execution time in the slow log, excluding lock waiting time.
This provides a more powerful tool for query performance diagnosis and optimization, facilitating the location of performance bottlenecks.
﻿
Aggregation and transformation
$convert: supports conversion between strings and BinData.
$toUUID: provides a simplified syntax for converting strings to UUID.
This enhances data processing and conversion capabilities.
﻿
Operation control
The defaultMaxTimeMS parameter is added to allow settings of the default timeout period on the server side for operations such as find and aggregate.
This helps prevent long-running queries from exhausting resources and improves system stability.
Security feature
Queryable encryption
It is supported to use range query operators (such as $lt, $lte, $gt, and $gte) on the encrypted fields.
This provides more complex query capabilities while ensuring data privacy, thereby enhancing security.
Ops improvement
Logging improvement
Starting from MongoDB 8.0, the database profiler can be configured to log slow operations based on the actual processing time spent by MongoDB (instead of the total latency of the slow operation). For more information, see Logging.
This avoids misclassifying operations as slow operations due to lock waiting or traffic control.
﻿
Index building enhancement
Fast error reporting: Index building errors are returned immediately during the scanning phase without the need to wait for completion.
Elastic deployment: Secondary nodes can request to stop problematic index building to prevent crashes.
Disk space management: Index building is automatically stopped when the available disk space falls below the value of indexBuildMinAvailableDiskSpaceMB.
This makes the index building process more robust, enables faster fault recovery, and minimizes the impact on businesses.
﻿
Batch and concurrent operations
bulkWrite: performs insertion, update, and deletion operations on multiple collections in a single request.
Concurrent DDL: supports concurrent execution of DDL operations on different collections under the same database.
This enhances the Ops efficiency and batch data processing capabilities.
﻿
Admission control
Entry queue: controls network request queuing through the ingressAdmissionControllerTicketPoolSize parameter.
This provides a new traffic control mechanism to prevent system overload.
﻿
Query settings
setQuerySettings: allows specifying indexes or rejecting specific query shapes.
explain() enhancement: adds queryPlanner.optimizationTimeMillis to display optimization time.
This provides more fine-grained control and insights into the query behavior. TencentDB for MongoDB supports more flexible traffic throttling policies. For details, see SQL Traffic Throttling.

Was this page helpful?

You can also Contact Sales or Submit a Ticket for help.

Yes

Feedback

Category	Feature Name	Description	Use Case/Impact
Performance optimization	Enhanced TCMalloc	Caching mechanism: changed from per-thread caching to per-CPU caching to reduce memory fragmentation and improve adaptability under high load. Memory release: a background thread that attempts to release memory to the operating system per second is added. The tcmallocReleaseRate parameter allows precise control of the release rate (in the unit of Bytes/s), with a default value of 10 MB/s.	More efficient memory usage is achieved, and it is especially suitable for scenarios with high memory pressure and significant load fluctuations.
		Majority write optimization	When write concern is set to "majority", acknowledgment is returned once the oplog is written to a majority of nodes, without waiting for the data application.	This significantly improves throughput and reduces latency for majority writes.
		oplog processing optimization	MongoDB 8.0 introduces an oplog buffer on secondary nodes, decoupling the data synchronization process into two independent stages: the Writer thread continuously pulls and buffers oplogs, while the Applier thread asynchronously applies data from the buffer. This allows parallel execution of the two threads, eliminating resource waste caused by serial waiting.	This improves the data replication throughput of secondary nodes and reduces replication latency between primary and secondary nodes.
		Batch insertion optimization	Non-transactional batch insertions are merged into a single oplog entry, with all documents sharing the same clusterTime in the Change Stream.	This improves batch insertion performance and reduces replication latency caused by secondary nodes replaying a large volume of oplog entries.
Shard enhancement	reshard optimization	`forceRedistribution`: allows for resharding using the same shard key to quickly redistribute data to new shards. `commitReshardCollection`: enables proactive disabling of write operations during the Ops window to accelerate the completion of data shard migration.	This provides faster and more flexible data redistribution, which facilitates load balancing and capacity planning.
		Shard management commands	`unshardCollection`: unshards a collection by merging data into a specified shard. `moveCollection`: moves an unsharded collection to any shard (excluding time series and encrypted collections). `shardAndDistributeCollection`: redistributes data immediately after sharding.	This provides more granular and flexible collection-level data layout management capabilities.
		Shard operation optimization	Hashed sharding: By default, 1 chunk is created per shard (previously 2 chunks per shard). findAndModify/deleteOne: Partial shard keys can be used as query conditions. updateOne with upsert: For sharded collections, the query conditions of the upsert operation do not need to include all the shard keys. directShardOperations role: Commands are allowed to be run directly on shards for shard operations.	This simplifies Ops and queries for sharded clusters, enhancing usability.
		serverStatus output enhancement	Starting from MongoDB 8.0, the serverStatus output includes the following new shardingStatistics fields: shardingStatistics.countTransitionToDedicatedConfigServerStarted shardingStatistics.countTransitionToDedicatedConfigServerCompleted shardingStatistics.countTransitionFromDedicatedConfigServerCompleted shardingStatistics.configServerInShardCache	This provides a more comprehensive view of the running status and internal statistical metrics of sharded clusters.
Query and aggregation	Query statistics and optimization	$queryStats: returns detailed statistics for recorded queries. Query Shape: introduces a new query shape, upgrading query pattern recognition from a passive cache management mechanism to an independently configurable and manageable resource. workingMillis: displays actual execution time in the slow log, excluding lock waiting time.	This provides a more powerful tool for query performance diagnosis and optimization, facilitating the location of performance bottlenecks.
		Aggregation and transformation	$convert: supports conversion between strings and BinData. $toUUID: provides a simplified syntax for converting strings to UUID.	This enhances data processing and conversion capabilities.
		Operation control	The defaultMaxTimeMS parameter is added to allow settings of the default timeout period on the server side for operations such as find and aggregate.	This helps prevent long-running queries from exhausting resources and improves system stability.
Security feature	Queryable encryption	It is supported to use range query operators (such as $lt, $lte, $gt, and $gte) on the encrypted fields.	This provides more complex query capabilities while ensuring data privacy, thereby enhancing security.
Ops improvement	Logging improvement	Starting from MongoDB 8.0, the database profiler can be configured to log slow operations based on the actual processing time spent by MongoDB (instead of the total latency of the slow operation). For more information, see Logging.	This avoids misclassifying operations as slow operations due to lock waiting or traffic control.
		Index building enhancement	Fast error reporting: Index building errors are returned immediately during the scanning phase without the need to wait for completion. Elastic deployment: Secondary nodes can request to stop problematic index building to prevent crashes. Disk space management: Index building is automatically stopped when the available disk space falls below the value of indexBuildMinAvailableDiskSpaceMB.	This makes the index building process more robust, enables faster fault recovery, and minimizes the impact on businesses.
		Batch and concurrent operations	bulkWrite: performs insertion, update, and deletion operations on multiple collections in a single request. Concurrent DDL: supports concurrent execution of DDL operations on different collections under the same database.	This enhances the Ops efficiency and batch data processing capabilities.
		Admission control	Entry queue: controls network request queuing through the ingressAdmissionControllerTicketPoolSize parameter.	This provides a new traffic control mechanism to prevent system overload.
		Query settings	setQuerySettings: allows specifying indexes or rejecting specific query shapes. explain() enhancement: adds queryPlanner.optimizationTimeMillis to display optimization time.	This provides more fine-grained control and insights into the query behavior. TencentDB for MongoDB supports more flexible traffic throttling policies. For details, see SQL Traffic Throttling.

tencent cloud