On this page
Serverless instances are in preview and do not support this feature at this time. To learn more, see Serverless Instance Limitations.
Atlas moves infrequently accessed data from your Atlas cluster to a MongoDB-managed read-only Data Lake on a cloud object storage. Once Atlas archives the data, you have a unified view of your Atlas and Online Archive data in a read-only Data Lake.
Atlas archives data based on the criteria you specify in an archiving rule. The criteria varies based on the type of collection you want to archive:
Online Archive in Atlas is available only on
greater clusters that run MongoDB 3.6 or later.
To create an Online Archive, you must have one of these roles:
How Atlas Archives Data¶
To archive data:
Atlas periodically runs a query to determine the documents that match the criteria for archiving. Atlas refers to this query as a job. Initially, Atlas runs the job every five minutes. If the size or number of documents to archive doesn't meet the threshold, Atlas expands the job interval by ten minutes, up to a maximum of twelve hours. If the job interval reaches the maximum or if either the size or number of documents to archive reaches the threshold, Atlas runs the job again and resets the job interval to five minutes.Note
The threshold is 2GB per job.
Atlas runs an index sufficiency query to determine the efficiency of the archival process. If the number of documents scanned to the number of documents returned is 10 or more, the query result triggers an
Index Sufficiency Warning. This warning indicates that you have insufficient indexes for an efficient archival process. For date-based archives, you must index the date field. For custom criteria that use an expression, Atlas might first convert a value before it evaluates it against the query.
For documents that match the archival criteria, Atlas:
- Writes to up to a maximum of 10,000 partitions per archival job.
- Writes up to 2GB of document data to partitions on the cloud object storage for each unique combination of query field values except dates, which are grouped during each run to reduce the number of partitions.
- Writes each subsequent quantity of document data (up to 2 GB) with each query run.
Online Archive runs on your Atlas cluster and utilizes the same underlying resources (e.g. IOPS). The default limit of 2GB per job prevents the operation from utilizing too many resources. If your cluster is currently satisfying workloads at the edge of its resource limits, you could push it past its capacity by activating Online Archive. Ensure that your Atlas cluster has excess resources before activating Online Archive.
If you activate Online Archive for an AWS cluster, the cloud object storage exists in the same region in AWS as your cluster. If you activate Online Archive for a Google Cloud or Azure cluster, Online Archive creates the archive in the AWS region closest to your cluster's primary based on a calculation of distance between the cluster and cloud object storage.
Atlas encrypts your archived data using Amazon's server-side encryption S3-managed keys (SSE-S3) for archived data. It can't use any encryption-at-rest encryption keys you might have used on your cluster data.
When you archive data, Atlas first copies the data to the cloud object storage and then deletes the data from your Atlas cluster. WiredTiger does not release the storage blocks of the deleted data back to the OS for performance reasons. However, Atlas eventually automatically reuses these storage blocks for new data. This helps the Atlas cluster to avoid fragmentation. To learn more about reclaiming the disk space, see How do I reclaim disk space in WiredTiger?.
Atlas provides a unified endpoint. You can use it to query both your live cluster and archived data using the same database and collection name that you use in your Atlas cluster. You can't use the unified endpoint over a Network Peering Connection, but you can set up a private endpoint or use a standard internet connection over TLS.
Data Lake for Online Archive¶
When you configure your
M10 or greater Atlas cluster for
Online Archive, Atlas creates a read-only Data Lake,
one per cluster, on a cloud object storage for your archived data.
Online Archive Data Lake can't perform the following activities:
Viewing the Online Archive¶
To view your Data Lake for the Online Archive:
- Navigate to the Data Lake page in the Atlas console.
- Click Data Lake from the left navigation in your Project page.
Querying the Online Archive¶
To query your Online Archive data, use the connection string through the Data Lake Connect button to connect to the cloud object storage.
You can also query your Online Archive data with SQL. For more information, see Querying with SQL.
Editing Online Archives¶
Once Atlas creates the Online Archive, you can't change the archiving criteria from Date Match to Custom Filter, or vice versa.
Deleting Online Archives¶
If you delete all the Online Archives, Atlas deletes the Data Lake. After deleting all the Online Archives, if you create an Online Archive with the same settings as a deleted Online Archive, Atlas creates a new Data Lake for the new Online Archive.
Online Archive Costs¶
Online Archive stores infrequently accessed data to lower the data storage costs on your Atlas cluster. You incur costs for the following items:
- Storage on the cloud object storage
- Data processed for queries
- Data returned by queries
To learn more about Data Lake costs, see the billing documentation for Atlas Data Lake.
Manage Your Online Archive¶
You can configure an Online Archive for a collection on your cluster through your Atlas console and API. Once you create an Online Archive, you can: