Customize Cluster Storage

Low CPU Class

Generally, the Low CPU option offers a lower-cost alternative for applications that require more memory and fewer CPUs. This option includes half the vCPUs of an instance with the General option of the same tier. For example, a General M40 instance includes 4 vCPUs, while a Low CPU M40 instance includes 2 vCPUs. Depending on the cluster tier, this option may also include fewer max connections. To learn more, see Connection Limits and Cluster Tier.

NVMe Storage

You can select the Local NVMe SSD storage option for some dedicated clusters that run on AWS or Azure. Locally attached ephemeral NVMe SSDs offer the highest level of speed and performance.

File Copy Based Initial Sync is enabled for dedicated clusters with NVMe running on MongoDB 6.0+.

Oplog Size Behavior

Atlas manages the cluster's oplog size and its entries differently, depending on whether storage auto-scaling is enabled for the cluster.

• Atlas enables cluster storage auto-scaling by default. If you choose to use storage auto-scaling, Atlas manages the oplog entries based on the minimum oplog retention window (oplogMinRetentionHours) setting. The oplog entries are time-stamped. The oplog window is the time difference between the newest and the oldest timestamps in the oplog.

  By default, Atlas sets the minimum oplog retention window to 24 hours. This means that, unless you set the minimum oplog retention window to a particular custom value, the mongod for the Atlas cluster retains all oplog entries for at least 24 hours and until the oplog reaches the maximum size that MongoDB best practices permit.

  You can set the minimum oplog retention window to a particular value in the Atlas UI. For storage auto-scaling to complete successfully, Atlas requires the minimum oplog retention window (oplogMinRetentionHours) up to (60 seconds) * (GB of disk space configured).

• You can opt out of cluster storage auto-scaling by un-checking the Storage Scaling checkbox in the Auto-scale section. If you opt out of storage auto-scaling, Atlas manages the oplog size as follows:

  • If you don't specify the oplog size in the UI, Atlas sets the oplog size to:

    • 5% of the disk size when you create a cluster, for General and Low-CPU clusters

    • 10% of the disk size for clusters with NVMe storage.

    Atlas doesn't change the oplog size if you change the storage size.

  • You may choose to scale up the oplog size when you scale up the cluster's storage. In this case, manually set the oplog size to a particular value when you create a cluster. As you increase the cluster's storage, Atlas scales the oplog size as follows:

    • For General and Low-CPU clusters, the oplog size scales up to remain at 5% of the storage capacity, not to exceed a certain maximum determined according to MongoDB best practices.

    • For clusters with NVMe storage, the oplog size scales up to remain at 10% of the storage capacity, not to exceed a certain maximum determined according to MongoDB best practices.

  If the cluster's storage capacity decreases, Atlas doesn't change the oplog size.

Change Storage Capacity or IOPS on AWS

Atlas handles changes differently based on whether you want to increase or decrease storage capacity or storage throughput (IOPS) on AWS.

Change Storage Capacity on Azure

For clusters deployed on Azure, you can change storage capacity in preset amounts only. To change server data volume size:

• Specify your preferred disk size in the text box.

• Move the slide bar until the text box displays your preferred disk size.

When you change a cluster's storage capacity, Atlas modifies the size of the servers' data volumes in a rolling manner with zero downtime. Atlas increases the storage capacity in place without copying data or performing an initial sync.

However, Azure doesn't allow in-place storage capacity downgrades. If you downgrade the storage capacity of a cluster, Atlas replaces each node and performs an initial sync for each node in the cluster.

During this time, you can still access your cluster, but each node that Azure modifies remains unavailable until the initial sync for that node completes.

Before you apply your storage capacity changes, the Review Changes page notifies you that Atlas triggers a rolling restart of your cluster when you make this change.

Change Storage Capacity on Google Cloud

When you change a cluster's storage capacity, Atlas modifies the size of the servers' data volumes in a rolling manner with zero downtime.

Atlas increases the storage capacity in place without copying data or performing an initial sync. However, Google Cloud doesn't allow in-place storage capacity downgrades.

If you downgrade the storage capacity of a cluster, Atlas provisions new volumes, and then syncs data from the old to the new volumes.

Before you apply your storage capacity changes, the Review Changes page notifies you that Atlas triggers a rolling restart of your cluster when you make this change.

Change Storage Capacity for Multi-Cloud Provider Clusters

Atlas selects the lowest common denominator across the three cloud providers. This ensures consistency across the multi-cloud deployment.

The following limitations might apply:

• You can't adjust the IOPS for your multi-cloud cluster.

• If your multi-cloud cluster includes Azure, you can change storage capacity in preset amounts only. See Change Storage Capacity on Azure.

Provisioned IOPS vs Standard IOPS in AWS

Provisioned IOPS let you customize the maximum IOPS rate for your cluster. They also:

• Deliver their configured IOPS rate more consistently when compared to standard IOPS.

• Lower your cluster's p90 latency (measurement of the server's response time). 90 percent of server requests have responses faster than the p90 latency value, so a lower p90 latency value means a generally faster response time.

To learn more about the merits of using provisioned vs standard IOPS, see Amazon EBS-optimized instances. See the following summary:

• General Purpose SSD volumes are designed to deliver their baseline performance 99% of the time.

• Provisioned IOPS SSD volumes are designed to deliver their provisioned performance 99.9% of the time.

Configure the IOPS Rate

Minimum Disk Capacity to RAM Ratios

Atlas enforces the following minimum ratios for given cluster tiers. This keeps cluster performance consistent with large datasets.

Instance sizes M10 to M40 have a ratio of disk capacity to system memory of 60:1. Instance sizes greater than M40 have a ratio of 120:1.

Atlas has disk capacity limits on single replica sets, scaling up to 4 TB for higher cluster tiers. To expand the total cluster storage beyond default limits, you can enable extended storage in the Project Settings. To accommodate further scaling in the future, we recommend that you enable sharding for long-term expansion.