Dynatrace periodically deletes transaction storage, Session Replay storage, and Log monitoring storage data that is older than the configured retention time. Adaptive Data Retention is a functionality according to which Dynatrace periodically increases or decreases the retention time of this data if disk space is insufficient.
Each data type stored on a disk has a default retention time, which specifies how long that data type can be stored on disk.
You can view the current environment storage quotas and retention times specific for your environment at Cluster Management Console > Environments.
The maximum amount of a given data type that can be stored on disk is determined as follows:
The configured retention time defines how long data of a certain type should be stored on the disk. Older data is periodically deleted.
The adapted retention time is the configured retention time reduced by a factor calculated when the following conditions apply:
If neither of these conditions apply, the adapted retention time is equal to the configured retention time.
For example, suppose all of the following statements are true:
In this case, the adapted retention time would be 9 days for session storage data:
Real retention time = 10 days × ((100 − 10) 100) = 10 × 0.9 = 9 days
This would result in all Session Replay storage data older than 9 days being deleted immediately, even though the configured retention time is 10 days.
Retention time is decreased if:
These conditions are continuously evaluated. If one of these conditions is violated, retention time is decreased and deletion of data older than the adapted retention time begins.
After data is deleted according to the adapted retention time:
If retention times are currently adapted, you should see a note on the Cluster Management Console > Environments > [environment] page, section Storage settings, under the affected data type.
Additionally, you'll receive a cluster event. For example:
The node with id 123 is not properly configured
The disk volume 'dynatrace-storage' attached to the cluster node 123 is too small to persist the expected amount of data in the cluster. This cluster node will be shut down. Make sure that all cluster nodes have the same partition size for the 'dynatrace-storage'. After the resize operation, restart the cluster node using the 'dynatrace.sh' command. To troubleshoot, visit: Storage recommendations.
Adaptation of data retention time can be caused by an undersized disk or an inappropriate quota.
To prevent unwanted adaptation of retention time for data:
Adaptive Data Retention is impacted by several factors. Under certain rare circumstances, the adaptation of retention time can lead to excessive data deletion for an environment. As a result, the amount of processed data that is retained and the amount of available storage, which is limited by disk space or quota, are not aligned.
To mitigate this issue:
Because retention time adaptation is a percentage of your data retention time, a longer data retention time proportionally causes more data—and therefore more hours—to be deleted.
As the new environment takes up disk space, the existing environments' retention times are gradually decreased. After some time, all of the cluster's environments have the same percentage change of their configured data retention times.
Some deviation is expected. When the deviation is too extreme, the required disk space for the configured retention time usually diverges heavily from the available disk space. Try setting the retention times to a more realistic value better suiting the currently available disk space.
The Adaptive Data Retention reduction displayed in Cluster Management Console is only a snapshot of the current situation. After the data is removed from the disk, the reduction is decreased continuously, resulting in this behavior.