Parameter Name | Restart Required After Modification | Default Value | Reference Value | Supported Version | Supported Instance Type | Applicable Scope | Parameter Description |
operation.profiling.slowOpThresholdMs | No | 100 | [0-65536] | 4.0,4.2,4.4,5.0,6.0,7.0 | Replica and shard | mongod,mongos | Sets the threshold time (in milliseconds) for determining a slow log query. |
operationProfiling.mode | No | off | [off | slowOp | all] | | Replica and shard | mongod | This parameter configures the database operation performance profiling mode. off: The performance analyzer is disabled and does not record any operations. slowOp: Only "slow operations" whose execution time exceeds the slowOpThresholdMs threshold (100 ms by default) are recorded. all: All database operations are recorded. |
setParameter.cursorTimeoutMillis | No | 600000 | [1-2147483647] | | Replica and shard | mongod,mongos | This parameter sets the maximum lifetime for a server-side cursor when it is inactive. The cursor is automatically closed upon timeout to release resources. Unit: milliseconds. By default, the value of this parameter is 10 minutes, or 600,000 milliseconds. If you set this parameter to 0, the cursor timeout mechanism is disabled. The cursor will remain enabled until the client explicitly disables it. |
setParameter.intenalQueryExecMaxBlockingSortBytes | No | 33554432 | [33554432-268435456] | | Replica and shard | mongod | This parameter sets the maximum amount of memory (in bytes) that a single sort operation can use. If this limit is exceeded, the operation will switch to slower disk-based sorting. |
setParameter.maxTransactionLockRequestTimeoutMillis | No | 5 | [0-60] | | Replica and shard | mongod | Sets the maximum wait time (in milliseconds) for a transaction to obtain a lock resource. If the lock is not obtained within this time, a timeout exception is thrown. |
setParameter.transactionLifetimeLimitSeconds | No | 60 | [5-300] | | Replica and shard | mongod | Sets the maximum duration (in seconds) for a single transaction from its start to automatic termination. |
balance.window | No | NULL | [00:00 | 23:00] | | Shard | mongos | Configures the time window for a sharded cluster to perform data balancing operations. The format is "start_time-end_time" (for example, "00:00-06:00"). |
openBalance.window | No | false | [true | false] | | Shard | mongos | Enables or disables the balance window. |
setParameter.tcmallocAggressiveMemoryDecommit | No | true | [true | false] | | Replica and shard | mongod,mongos | Specifies whether to enable fast memory reclamation. |
cmgo.crossZoneLoadBalancing | No | off | on|off | | Shard | mongos | Configures whether to evenly distribute read and write requests across multiple Availability Zones (AZs). |
cmgo.retainVipWhenSwitchover | No | on | on|off | | Replica and shard | mongos | Controls whether the Virtual IP (VIP) remains unchanged during a primary-secondary switchover/failover. |
setParameter.ttlDeleteBatch | No | 2147483647 | [1-2147483647] | | Replica and shard | mongod | Controls the batch size used by Time-To-Live (TTL) indexes when expired documents are deleted. Note: In MongoDB versions earlier than 6.0, the TTL index background thread scans for expired data and deletes it every 60 seconds. If a large number of expired documents are identified in the current cycle, the delete operations are performed intensively. This generates significant write load and lock contention, leading to increased latency for normal business requests and periodic jitter in the instance. It is recommended to properly control the number of documents cleaned up in a single operation. |
setParameter.ttlMonitorSleepSecs | No | 60 | [60-600] | | Replica and shard | mongod | Controls the execution frequency of the MongoDB TTL monitor thread (TTLMonitor), specifically the number of seconds the thread sleeps between each run. |
setParameter.failIndexKeyTooLong | No | true | [true | false] | 3.2,3.6,4.0 | Replica and shard | mongod | Controls MongoDB's validation behavior for index key length, determining whether to force a failure when the index key length exceeds the system limit. Enable (true): The index key length limit is strictly enforced, and length validation is performed during index creation. Disable (false): The length limit is relaxed, allowing the creation of extra-long index keys (which may be truncated). |
storage.wiredTiger.collectionConfig.blockCompressor | Yes | snappy | [snappy | zlib | zstd] | 4.2,4.4,5.0,6.0,7.0 | Replica and shard | mongod | Specifies the block compression algorithm used by the WiredTiger storage engine for collection data in MongoDB. snappy (default): Balance compression ratio and CPU overhead, making it suitable for most scenarios. zlib: Provide a higher compression ratio at the cost of significantly increased CPU usage. zstd: Save 20–30% more disk space than snappy (depending on data type); offers decompression speed close to snappy and significantly faster compression than zlib. |
setParameter.migrateCloneInsertionBatchDelayMS | No | 0 | [0-5000] | | Shard | mongod | This parameter controls the wait time (in milliseconds) after each batch of documents is inserted into the destination shard during data migration in a MongoDB sharded cluster. It is used to moderate the I/O pressure on the destination shard. |
setParameter.rangeDeleterBatchDelayMS | No | 20 | [0-5000] | | Shard | mongod | This parameter controls the delay interval (in milliseconds) between each batch deletion of old chunks from the source shard after data migration in a MongoDB sharded cluster. It is used to regulate the I/O pressure caused by the deletion operations on the source shard. |
setParameter.rangeDeleterBatchSize | No | 128 | [0-2147483647] | | Shard | mongod | This parameter controls the upper limit on the number of documents deleted per batch from the source shard after data migration in a MongoDB sharded cluster. It is used to balance deletion efficiency and cluster load. |
setParameter.migrateCloneInsertionBatchSize | No | 0 | [0-5000] | | Shard | mongod | This parameter controls the upper limit on the number of documents inserted per batch into the destination shard during data migration in a MongoDB sharded cluster. It is used to balance migration speed and the load on the destination shard. |
setParameter.internalQueryMaxBlockingSortMemoryUsageBytes | No | 33554432 | [33554432-268435456] | | Replica and shard | mongod,mongos | This parameter controls the maximum size (in bytes) of the in-memory sort buffer that a single MongoDB query can use when performing a blocking sort, which occurs when an index is not available. If this limit is exceeded, MongoDB will sort using temporary files on disk. |
피드백