Time Series Archive Rules

Parent: Archive Rules · Published reference · snapshot 2026-10-15

Depth-first rabbithole dossier for Time Series Archive Rules; source-anchored research pack.

These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.

Definitions

[low] Online Archive is not a backup or point-in-time restore substitute, and if historical access depends on unsupported stages, that path must be redesigned before enabling the feature. Sources: https://dev.to/dvorlandi/atlas-online-archive-for-large-collections-keep-hot-data-hot-33fm [source]

How it works

[high] For a time series archive, the date field value must be in ISODate format. The Terraform date_format values (ISODATE default, EPOCH_SECONDS, EPOCH_MILLIS, EPOCH_NANOSECONDS, OBJECT_ID) are documented generally; the sources say only that type must be DATE when collection_type is TIMESERIES. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md, https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/ [source]
[low] When Online Archive archives a time series collection, it sets the internal `control.closed` field to `true` on each bucket marked for archival, preventing writes to that bucket. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[high] MongoDB introduced the Online Archive feature from Atlas 4.4 onward to manage data lifecycle; Atlas evaluates archiving rules continuously and transfers matching documents to cloud object storage. For date rules, the Terraform docs state data is archived when the current date exceeds the date field value plus the number of days. (verified as of 2026-10-15) Sources: https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/, https://www.mongodb.com/docs/atlas/online-archive/overview.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-ti [source]
[low] The `BucketCatalog` closes a bucket if a new measurement would cause it to span more time than the collection settings allow (such buckets remain eligible for reopening), and archives a bucket if a new measurement's timestamp is earlier than the minimum timestamp for the current open bucket. Sources: https://raw.githubusercontent.com/mongodb/mongo/master/src/mongo/db/timeseries/README.md [source]

Problems, failure modes and limitations

[high] Once an Online Archive is created, the namespace, storage region, date field, and time field are immutable and cannot be modified. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md, https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/ [source]
[high] An Online Archive must have at least one partition field, with values limited to 700 characters; documents exceeding this limit are not archived. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md, https://raw.githubusercontent.com/mongodb/terraform-provider-mongodbatlas/master/docs/resources/online_archive.md, https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/, https://marketplace.upbound.io/providers/crossplane-contrib/provi [source]
[low] An error occurs when starting an archive if the archive collection type does not match the collection type found. Sources: https://github.com/mongodb/terraform-provider-mongodbatlas/issues/1081 [source]

Facts and statements

[low] Create a compound index on `[control.closed, control.min.<date_field>]` for the underlying `system.bucket.<collection>` to ensure optimal performance for time series archives. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md [source]
[low] Online Archive for Time Series is currently available only if the cluster is on AWS, and the checkbox for this feature is not visible if the cluster is on Azure or GCP. (verified as of 2026-10-15) Sources: https://www.mongodb.com/community/forums/t/mongodb-atlas-online-archive-for-timeseries-collection/256652 [source]
[low] If an online archive is deleted before buckets are removed from the cluster, the buckets remain, and delete operations on the time series collection return a `deletedCount` of 0 for documents in those buckets. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[low] Indexes offer better performance during archival, and you can modify, pause, and delete online archival rules at any time. Sources: https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/ [source]
[low] For MongoDB versions earlier than 8.2, run the delete on the `system.buckets.<collection>` collection; for 8.2 and later, run the delete command on the time series collection with the `rawData: true` option. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[low] Query-based reopening relies on a `{<metaField>: 1, <timeField>: 1}` index, which is created by default for new time series collections in v6.3+; if the index does not exist, query-based reopening is not used. (verified as of 2026-10-15) Sources: https://raw.githubusercontent.com/mongodb/mongo/master/src/mongo/db/timeseries/README.md [source]
[high] Online Archive is supported only on dedicated clusters of tier M10 and above, and is not supported on Atlas Infinite clusters, Flex clusters, or Atlas serverless clusters during public preview. (verified as of 2026-10-15) Sources: https://www.mongodb.com/docs/atlas/online-archive/overview.md, https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/, https://dev.to/dvorlandi/atlas-online-archive-for-large-collections-keep-hot-data-hot-33fm, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas [source]
[high] Configuring an Online Archive creates two federated database instances: one for querying archive data only and one for querying both cluster and archive data, accessible via separate federated connection strings. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md, https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/, https://dev.to/dvorlandi/atlas-online-archive-for-large-collections-keep-hot-data-hot-33fm [source]
[low] Data Federation supports queries across Atlas clusters, online archives, Data Lake datasets, AWS S3 buckets, and HTTP stores, allowing read-only consumption of archived data by downstream environments to reduce main cluster traffic. Sources: https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/ [source]
[medium] Atlas determines the data type of a partition field by sampling 10 documents, and will not archive a document if its specified field value does not match values in other documents in the same collection. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md [source]
[medium] Five days after deleting an online archive, Atlas automatically deletes all files and metadata on the storage, making the data unrecoverable. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md, https://www.mongodb.com/docs/atlas/online-archive/overview.md [source]
[low] If you drop a database or collection configured for online archive, the archived data continues to be available, but if you delete all online archives, the federated database instance is deleted and recreated upon new archive creation. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[low] To find partitions, Atlas runs operations on the archive that find up to 1000 partitions per operation, executing the minimum number of operations required to satisfy the query. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[low] Blocking queries have performance characteristics associated with the slowest storage (archive), while streaming queries like finds have performance characteristics associated with the highest performing storage (Atlas cluster), returning results as soon as available. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[high] Capped collections do not support online archival, and writing to the Online Archive is not supported. Sources: https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/, https://www.mongodb.com/docs/atlas/online-archive/overview.md [source]
[low] To create or delete an online archive, you must have the Project Data Access Admin, Project Cluster Manager, or Project Owner role. Sources: https://foojay.io/today/atlas-online-archive-efficiently-manage-the-data-lifecycle/ [source]
[high] The default timeout for creating an online archive is 3 hours, and the flag to wait for the state to reach IDLE or ACTIVE defaults to false. Sources: https://raw.githubusercontent.com/mongodb/terraform-provider-mongodbatlas/master/docs/resources/online_archive.md, https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/ [source]
[high] Valid online archive states include Pending, Archiving, Idle, Pausing, Paused, Orphaned, and Deleted. Sources: https://raw.githubusercontent.com/mongodb/terraform-provider-mongodbatlas/master/docs/resources/online_archive.md, https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/ [source]
[low] AZURE may be selected as the cloud service provider for archived data only if Azure is the cluster's provider and no AWS online archive has been created for the cluster. Sources: https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/ [source]
[low] MongoDB stores time-series data in bucket documents, and Online Archive operates on full bucket documents only to preserve column compression efficiency. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md [source]
[low] A bucket is permanently closed by setting the `control.closed` flag, which makes it ineligible for reopening, and this can only be done manually for Atlas Online Archive. Sources: https://raw.githubusercontent.com/mongodb/mongo/master/src/mongo/db/timeseries/README.md [source]
[low] The `BucketCatalog` archives or closes idle buckets if memory usage exceeds the `timeseriesIdleBucketExpiryMemoryUsageThreshold`, removing most in-memory state but retaining a small record for quicker reopening. Sources: https://raw.githubusercontent.com/mongodb/mongo/master/src/mongo/db/timeseries/README.md [source]
[low] Online Archive deletes documents using `_id` and shard keys, so users must pause the archive before sharding operations and refrain from updating documents once they become eligible for archival to prevent data loss. Sources: https://www.mongodb.com/docs/atlas/online-archive/manage-online-archive.md [source]
[medium] If a cluster already has an `Active` online archive with the same archiving rule for the same database and collection, creating a new one fails; however, if the existing archive is `Paused` or `Deleted`, the new archive is created with `Active` status. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md [source]
[medium] Before configuring an Online Archive, ensure the Atlas cluster is healthy; if one node is down, the archive remains in `PENDING` status until the cluster is fully operational. Sources: https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-standard.md, https://www.mongodb.com/docs/atlas/online-archive/configure-online-archive-time-series.md [source]
[high] The state of the online archive is required for pausing an active archive or resuming a paused one, and the resume request fails if the collection has another active online archive. Sources: https://www.pulumi.com/registry/packages/mongodbatlas/api-docs/onlinearchive/, https://raw.githubusercontent.com/mongodb/terraform-provider-mongodbatlas/master/docs/resources/online_archive.md [source]

Related concepts

Archive — is a part of Time Series Archive Rules
Time — is a part of Time Series Archive Rules
Series — is a part of Time Series Archive Rules
Rules — is a part of Time Series Archive Rules

Children

No children recorded.

← the whole tree · 3D view· how to read this page