The amount of data that applications generate is growing exponentially, introducing cost and infrastructure complexity for many companies today. MongoDB Atlas makes it easy to manage your entire data lifecycle without replicating or migrating it across multiple systems.

With MongoDB Atlas Online Archive, you can seamlessly tier your data across fully managed databases and cloud object storage, all while retaining the ability to query it through a single endpoint. Create a rule to automatically archive infrequently accessed data from your live MongoDB Atlas clusters to fully managed cloud object storage and save on operational and transactional data storage costs.

Online Archive allows customers to automatically tier data across Atlas clusters and MongoDB-managed cloud object storage and use a unified endpoint to query that data from their applications.

Here’s how it works:

1. First, you need to write an archiving rule by providing a namespace (database and collection) and one of the following options:
   a. Date Match: Date field (within the documents, can be nested), and age limit (number of days past the date field when archival should begin)
   b. Custom Filter: Write a custom query (e.g. archive = true)

2. Next, you can choose up to two commonly queried fields in addition to the date field which will allow us to partition archived documents for optimal query performance.

3. Once you confirm the details of the online archive, Atlas begins archiving documents that match the rule to fully-managed cloud object storage.

4. You can then use the new connection string you are provided with, to query both your Atlas cluster and their Online Archive simultaneously. You are also provided with an archive-only connection string to query archival data independent from live cluster data.

Archived documents are deleted from the Atlas cluster and cannot be updated or deleted once archived.You can pause, edit rules, or delete the archive at any time.

Online Archive archival jobs run every five minutes. Documents that match the customer’s archive rule are stored in a temporary collection on the Atlas cluster, then archived off in files up to 100MB, and archiving up to 2GB total per 5 minute interval. This is done so that archival jobs don’t overwhelm the cluster’s resources.

2GBs every 5 minutes is the fastest that archiving will happen, if the archival job is consistently archiving less than 2GB of data every 5 minutes then the interval will gradually decrease to further reduce demand on the cluster.