Is there a backup strategy for QuickSight?

Question & Answer
1

When if we really mess up something using the API and want to get back to a specific state at a point in time. Is that possible?

3 Likes
2

Hi, are you referring to the entire account backup - including all dashbaords, datasets, users, group; or are you referring to back up a particular asset like ā€˜datasetsā€™? For example, we have versioning feature in dataset, using which you can perform rollback in UI to a previous version.

3

@DRK for dataset, we have versioning: Reverting datasets back to previous published versions - Amazon QuickSight
You can always go back to previous published version.

For analysis: if you accidentally delete it, you can restore it with this API: QuickSight ā€” Boto3 Docs 1.21.33 documentation
I would suggest always publish a dashboard as the ā€œbackupā€ of an analysis. So, you can ā€œsave asā€ a dashboard to be an analysis.

For dashboard, you can always re-publish from the analysis. The only thing is how to re-store the access permissions you set up on the old dashboard. I published a solution with SSM parameter and Lambda functions to apply and store the access permissions: Build a centralized granular access control to manage assets and data access in Amazon QuickSight | AWS Big Data Blog

In future, we will release assets as bundle which can natively support the backup/restore use case. Thanks.

2 Likes
4

Hello AWS Team,

We treat our QuickSight assetsā€”datasets, analyses, and dashboardsā€”as critical ā€œinfrastructure-as-codeā€ components. From my understanding, QuickSight likely stores its assets in DynamoDB under the hood, which opens the door for a seamless backup solution.

Would it be possible to offer a built-in, scheduled backup featureā€”ideally integrated with AWS Backupā€”that regularly snapshots all QuickSight assets (e.g., daily)? This would be invaluable for ensuring business continuity, simplifying disaster recovery, and aligning with best practices for data-protection across the AWS ecosystem.

Iā€™d love to hear your thoughts on whether this is on the roadmap or if thereā€™s any potential workaround we can explore in the meantime.

Thank you for your consideration!