DynamoDB Backup and Point-in-Time Recovery Guide
DynamoDB protects your data two ways. On-demand backups are full snapshots you take and keep until you delete them.
Point-in-time recovery (PITR) is continuous, automatic backup that lets you restore the table to any second within a rolling window.
The examples use an append-only audit log for a multi-tenant SaaS: one item per
event, partitioned by tenant (PK = TENANT#acme, SK = EVENT#<time>#<id>), as
set up in the Operations & Cost overview.
It's a compliance record, so a bad migration or an accidental bulk delete has to be recoverable to the moment before the mistake.
How does DynamoDB backup and point-in-time recovery work?
- PITR = continuous backup, restore to any second within a recovery period you set from 1 to 35 days.
- On-demand backups = manual full snapshots kept as long as you want, independent of PITR's window.
- Restores create a new table. You restore to a new name, verify, then repoint the app. The original is untouched.
- PITR is billed per GB-month of table size (table data plus local secondary indexes), not per restore point; see DynamoDB pricing.
| PITR | On-demand backup | |
|---|---|---|
| Restore granularity | Any second in the window | The moment it was taken |
| Retention | 1–35 days, rolling | Until you delete it |
| Billed on | Table + LSI size, per GB-month | Backup size, per GB-month |
| Starts | Once enabled, continuously | Each time you request one |
The problem: a mistake you can't undo in place
DynamoDB has no transaction log you can roll back and no "undo" on a write.
If a migration script rewrites every event's action field, or someone runs a
delete that's wider than intended, the table is now in the wrong state.
Without backups, the data is gone.
For an audit log, whose value is being a trustworthy record, "we can't get last Tuesday's events back" is a compliance failure.
PITR restores to any second; backups keep a fixed moment
Point-in-time recovery, once enabled, takes continuous automatic backups with
per-second restore granularity
(AWS docs).
You set the window with RecoveryPeriodInDays, from 1 to 35:
aws dynamodb update-continuous-backups \
--table-name audit-log \
--point-in-time-recovery-specification PointInTimeRecoveryEnabled=true,RecoveryPeriodInDays=35You can restore to any second between EarliestRestorableDateTime and
LatestRestorableDateTime, which is typically five minutes before now. A
restore can also target a different Region.
Decreasing the recovery period immediately discards restore points older than the new period. Shortening the window doesn't lower the price either, since PITR bills on table size.
Disabling and re-enabling PITR resets the recoverable start time, so you lose the prior continuous history.
PITR also gates the managed export to S3: the
export reads from the same continuous backups, so with PITR off,
ExportTableToPointInTime fails with
PointInTimeRecoveryUnavailableException.
On-demand backups are separate: full-table snapshots you create explicitly and keep indefinitely, useful for a pre-migration checkpoint or a compliance archive beyond PITR's 35 days:
aws dynamodb create-backup --table-name audit-log --backup-name audit-log-pre-migrationThe request is processed instantly, the backup is restorable within minutes, it consumes none of the table's throughput, and there's no limit on how many you keep.
On-demand backups don't guarantee causal consistency across items. AWS puts the skew between updates at "usually much less than a second", but a backup taken mid-write-burst is not a single frozen instant.
For a checkpoint you need to be exact, stop the writers (or let the burst drain)
before you call create-backup.
What a restore doesn't bring back
A restore recreates the table data and its indexes, not its operational wiring.
Per the AWS docs, you must re-configure on the restored table: auto scaling policies, IAM policies, CloudWatch metrics and alarms, tags, stream settings, TTL, deletion protection, and PITR itself.
Forgetting PITR is how the next incident becomes unrecoverable. Re-apply each setting with one call:
aws dynamodb update-continuous-backups --table-name audit-log-restored \
--point-in-time-recovery-specification PointInTimeRecoveryEnabled=true,RecoveryPeriodInDays=35
aws dynamodb update-time-to-live --table-name audit-log-restored \
--time-to-live-specification Enabled=true,AttributeName=expiresAt
aws dynamodb update-table --table-name audit-log-restored \
--stream-specification StreamEnabled=true,StreamViewType=NEW_AND_OLD_IMAGES \
--deletion-protection-enabledThen re-tag the table with aws dynamodb tag-resource, re-run the
register-scalable-target / put-scaling-policy pair from auto
scaling, and recreate the alarms.
If the table is defined in infrastructure-as-code, one deploy against the
restored table re-applies all of it. The
infrastructure code generator
turns the original table's describe-table output into Terraform, AWS CDK or
CloudFormation with its streams and deletion protection (and TTL, if you paste
the describe-time-to-live output too); add PITR, tags, auto scaling policies,
alarms and IAM yourself.
At restore time you can deliberately change some settings (billing mode, provisioned capacity, encryption keys), and you can exclude some or all secondary indexes, which makes the restore faster and cheaper if you can rebuild them later.
Scheduled backups with AWS Backup
DynamoDB's own on-demand backups have no scheduler. For recurring backups with retention, DynamoDB integrates natively with AWS Backup (AWS docs).
Backup plans give you scheduled backups with lifecycle rules (including cold-storage tiering), automatic cross-account and cross-Region copies for DR, an independent KMS key via the backup vault, and Vault Lock so no one can delete a backup before its retention ends.
AWS Backup's advanced DynamoDB features are on by default for accounts that had no backup vault before 21 November 2021. A backup taken by an AWS Backup plan or an AWS Backup on-demand job opts the account in; backups from an Organizations backup policy don't.
Check a Region with aws backup describe-region-settings; if DynamoDB is
false, run:
aws backup update-region-settings \
--resource-type-opt-in-preference DynamoDB=true \
--resource-type-management-preference DynamoDB=trueBackups AWS Backup creates (type AWS_BACKUP) can't be deleted from the
DynamoDB console; manage them in AWS Backup.
A worked example: recovering from a bad migration
A migration meant to add an expiresAt attribute instead overwrote action on
every event with an empty string at 10:00 UTC. PITR is on with a 35-day window,
so you restore to the last second before the migration ran, into a new table:
aws dynamodb restore-table-to-point-in-time \
--source-table-name audit-log \
--target-table-name audit-log-restored \
--restore-date-time 2026-06-24T09:59:59Z| step | result |
|---|---|
| restore PITR to 09:59:59 | new table audit-log-restored with correct actions |
| diff against live | confirm only the migration's rows differ |
| cut app over to restored | original left intact for forensics |
The corrupted table stays untouched while you verify the restore. You compare
the restored events against the live ones, confirm the action values are
back, re-apply the settings above, then repoint the app.
If the loss were a handful of items rather than a whole-table corruption, you could instead copy just the affected rows from the restored copy into the live table; see copy a DynamoDB table.
Do it in DynoTable
The Settings section of the Table settings dialog (the Settings button in a table's toolbar) covers both kinds of backup.
On-demand backups lists the table's backups, yours and DynamoDB's system backups, newest first, with name, time and size. A Last backup line names the newest available one. Backups made by AWS Backup are not listed.
Create backup takes a backup with a generated name. Its row shows Creating… and turns available without reopening the dialog. You can delete a backup you made once it is available; system backups show their expiry and can't be deleted.
If the table was deleted and re-created under the same name, the earlier table's backups are marked Earlier table with this name.
Point-in-time recovery turns on or off with a confirmation each way. While it is on, the section shows the restorable window and the Recovery period, which you set from 1 to 35 days. Shortening it asks first, because the older restore points are deleted; lengthening it applies directly.
Delete table says what you can restore from afterwards: the 35-day system backup when PITR is on, otherwise the backup that is kept, otherwise that none was found. You can't delete the table while one of its backups is still being created.
A bulk edit or delete that writes directly can't be reverted. With PITR on, its summary shows the time to restore the table to. With PITR off, tick Create a backup first: the run takes a backup and waits about a minute before it writes, so the backup holds none of its changes.
That backup is kept, and billed, until you delete it. Listing backups is free; creating or deleting one and changing the recovery period are on paid plans. DynamoDB Local has no backups.
The restore itself stays in the AWS console, CLI or AWS Backup. DynoTable
connects to the restored table like any other, so you can query the affected
tenant's events, confirm the action values are correct, and compare against
the live table before you cut over.

You can also export the recovered events for an offline compliance record; see export DynamoDB to CSV.
Pitfalls and next steps
- Enable PITR before you need it. It only protects from the moment it's on;
there's no retroactive recovery. Run the
update-continuous-backupscommand above on any table whose data you can't afford to lose. - Disabling PITR resets the window. Toggling it off and back on wipes the
continuous history. To keep less history, lower
RecoveryPeriodInDaysinstead. Denydynamodb:UpdateContinuousBackupsin IAM to everyone but your platform role, and take an on-demand backup before any planned change. - Deleting a PITR table leaves a safety net. DynamoDB keeps a
<table>$DeletedTableBackupsystem backup for 35 days at no cost, which you can restore like any other backup. - Restores aren't instant or free. A restore provisions a whole new table and takes time proportional to size; budget for the duration and the extra table.
- 35 days isn't archival. For longer retention, take on-demand backups, run an AWS Backup plan, or export to S3.
To see how backups fit with transactions, Streams, TTL, capacity and global tables, go back to the Operations & Cost overview.
Download DynoTable to connect to a restored table and verify your recovery before you trust it.


