Managing Autoscaling Volumes
Identify eligible volumes, activate and deactivate autoscaling, and monitor grow and shrink actions in the Datafy app.
Once AutoScaler is installed in your environment, you can start activating autoscaling on your volumes. This page walks through the day-to-day actions of working with autoscaling volumes - identifying which volumes are eligible, activating and deactivating autoscaling, and monitoring grow and shrink actions as Datafy optimizes your storage.
For a deeper look at how the autoscaling mechanism works under the hood, check out How AutoScaler Works.
Identifying Supported Volumes
To find which volumes in your account are eligible for autoscaling, open the All Volumes table in Fleet Manager and use the AutoScaling State filter. Volumes fall into one of three states:
Active - Autoscaling is enabled and Datafy is managing the volume. The AutoScale toggle is on.
Inactive - The volume is eligible for autoscaling but not currently managed. The AutoScale toggle is off and can be turned on.
Unsupported - The volume cannot be autoscaled. A gray label in the AutoScale column explains why. Common reasons include AutoScaler not being installed on the instance, the volume being a boot volume, or an unsupported filesystem or volume type. For the complete list of supported configurations, see Supported Infrastructure.
If you're not sure which volumes to autoscale first, start with the largest volumes or those with the lowest utilization - these offer the highest savings potential. Sort the All Volumes table by Original Size or Utilization to find them.
Activating Autoscaling
There are two ways to activate autoscaling on eligible volumes:
Manual Activation - To activate autoscaling on a single volume, find it in the All Volumes table and turn on the AutoScale toggle.
Rule-Based Activation - For environments with many volumes or dynamic behavior, you can define autoscaling rules that automatically activate autoscaling on volumes matching specific criteria, such as volume tags, instance IDs, cluster names, or node groups. This is the recommended approach for managing autoscaling at scale.
Once autoscaling is activated for a volume, AutoScaler begins to copy the data from the original EBS volume to new, right-sized Datafy-managed volumes (see How AutoScaler Works for more details).
The progress of the activaiton is indicated next to the AutoScale toggle of the volume, and in the optimization actions log in the Reports page. The time this takes depends on the amount of data on the volume and the volume's IOPS and throughput. Your applications continue running normally throughout - Datafy always gives precedence to your application's I/O over the activation process.
Datafy offers different migration configurations optimized for different volume sizes and usage patterns. Contact us to discuss the best configuration for your environment.
There is a minimum volume size for activating autoscaling (default 50 GiB). Volumes close to the minimum will not shrink significantly. You can check your account's configured minimum in your account settings.
Autoscaling Multiple Volumes
When autoscaling is activated for multiple volumes on the same instance - whether manually or through a rule - the volumes are copied sequentially, one at a time, to minimize the impact on your applications. Volumes on different instances are activated in parallel regardless of account size.
Each instance supports up to 8 autoscaling volumes at a time. If your instance has more than 8 volumes, we recommend selecting the largest among them. Additional volumes cannot be activated until one is deactivated.
Monitoring Autoscaling Activity
Once autoscaling is active, Datafy continuously monitors your volumes and performs grow and shrink actions as needed (see How AutoScaler Works for details on how these operations work). You can track this activity on the Reports page, and drill into individual volumes through the Volume Detail page.
Tracking Grow and Shrink Actions
Every grow, shrink, activation, and deactivation is logged in the Optimization Actions table on the Reports page. Each entry shows the operation type, the volume size before and after, and the start and end times. The "Type" column indicates the progress of ongoing actions, and the final result for actions that failed or were canceled in the middle (for example, a shrink aborted because utilization increased).
Datafy grow actions will grow the underlying storage up to the filesystem's reported size, but not beyond it.
Each action is also reflected in the Usage chart as a step change in the Size line — a drop after a shrink, a jump after a grow.
To track a specific volume, click its ID in the All Volumes table to open the Volume Detail page. The same Optimization Actions table and charts are available there, scoped to that individual volume.
Use the filters on the Reports page to narrow results by cluster, instance, volume, or tags, and set a date range or auto-refresh interval to follow active operations.
For details on how grow and shrink thresholds are determined and how to tune them, see AutoScaling Configurations.
Understanding Your Savings
The gap between the original size of your volumes and their current size is storage Datafy has reclaimed, for which you are no longer paying.
Current Savings
The Autoscaling Storage widget shows you the size of all of the volumes that are autoscaled - both their original and current sizes. This provides a snapshot of the current state of the autoscaling volumes in your account.
For a dollar value estimation, check out the Estimated Monthly Saving widget at the top of the Reports page. It shows an estimation of your current realized savings, alongside the full potential savings if all eligible volumes were autoscaled.
Both widgets reflect the current state of your account, and may fluctuate over time as volumes are created and deleted, or autoscaling is activated.
Savings Over Time
To better understand your savings over time, use the Usage chart in the Reports page (and in the details page of every volume). The chart displays three plot lines aggregated across all of your monitored volumes, according to the filters you currently have selected:
Used - Actual data across your volumes.
Size - Current provisioned capacity of the Datafy-managed volumes.
Original Size - Total volume size before autoscaling was activated.
The area between Size and Original Size represents your total storage savings.
Select the Original Size in the graph legend plot to hide it and zoom in on the actual size of your volumes.
The complementary Utilization chart shows the average disk usage percentage across monitored volumes over time. Volumes managed by Datafy will typically show higher utilization than before activation, since Datafy right-sizes the underlying storage to match actual usage.
Autoscaling Volume Lifecycle
While a volume is autoscaling, Datafy manages it automatically - no manual intervention is needed for day-to-day operations.
Datafy is designed to be as seamless as possible in the day-to-day management of your storage. However, there are a few things that behave diffferently for autoscaling volumes that you need to be aware of:
Filesystem Mounting - On EC2 instances, autoscaling will not be enabled for volumes that do not follow the best practice of using the filesystem UUID in the relevant
/etc/fstabentry.
Mounting a volume using the device path (e.g., /dev/sdf or /dev/nvme1n1) can result in failure to mount an autoscaling volume after reboot or mounting the volume to the wrong mountpoint, and is not supported.
IaC Reconciliation - If you manage your infrastructure with Terraform or other IaC tools, autoscaling volumes can cause drift and state reconciliation issues. See IaC Reconciliation for setup instructions.
Modifying Volumes - Changing volume properties such as size, IOPS, or throughput through the AWS Console or AWS API is not supported for autoscaling volumes.
Do not modify autoscaling volumes directly through AWS. This can cause unexpected behaviors in your volumes and is not supported.
If modifying volume properties is part of your regular workflow, contact us to discuss your use case.
Volume Snapshots - To take snapshots of autoscaling volumes use the Datafy API or the CSI integration. See Snapshots for more details.
Deactivating Autoscaling
To deactivate autoscaling, turn off the AutoScale toggle in the All Volumes table or on the Volume Detail page. Like when autoscaling is activated, the progress will be indicated in the AutoScale column and in the Reports page.
When deactivation completes, the new volume has a new AWS EBS volume ID, but it retains the original volume's tags and is attached to the same device path and mountpoint. The Datafy virtualization layer is removed and the Datafy-managed volumes are deleted. From this point on, the volume operates as a normal EBS volume, fully independent of Datafy.
A reboot is required after deactivation to ensure clean filesystem state and proper device mapping. Your data is accessible immediately after the copy completes, but the reboot finalizes the transition.
Re-activating autoscaling for a specific volume will be blocked until the instance has been rebooted, or the volume has migrated to another node (in K8s).
If you'd like to programmatically deactivate autoscaling for your volumes, use the deactivate-autoscaling endpoint.
Exceptions From Autoscaling Rules
When autoscaling is deactivated for a volume, it is automatically marked as an exception to all autoscaling rules. This prevents any rule from re-activating autoscaling on that volume without your explicit action. If you want a rule to apply to the volume again in the future, you can remove the exception through the rule's detail panel.
Last updated
Was this helpful?