cloud-nuke v0.52.0 - Parallel Scanning and Deletion

   |   3 minute read   |   Using 556 words

The cloud-nuke project released version v0.52.0 on June 13, 2026, delivering a significant performance improvement for engineers cleaning up large AWS environments. This update introduces parallel resource scanning and deletion, allowing operators to clear out stale infrastructure much faster than before.

The full release notes and downloads are on the GitHub release page.

Parallel scanning and nuking

The headline feature of this release is the introduction of concurrent operations during both the discovery and destruction phases. Historically, cloud-nuke processed many resources in a more sequential fashion, which could lead to long wait times when dealing with hundreds of resource types across multiple regions.

A new CLI flag called --parallelism now controls the global concurrency cap. It defaults to 10, but operators can tune this value based on their account limits and the volume of resources they need to remove. Resource discovery now happens concurrently across different regions and resource types.

The deletion logic has also been optimized. Regional resources are now nuked concurrently while still respecting the internal dependency graph. For example, the tool ensures that resources within a region are deleted in the correct order to avoid dependency errors. Global resources, such as IAM roles, Route53 records, and CloudFront distributions, are handled differently. To ensure safety and consistency, the tool deletes these global entities last after all regional resources are gone.

Robust service linked role deletion

This release includes a specific fix for handling AWS service linked roles. These roles are often created by AWS services and require a specific deletion task to be initiated. A common issue occurred when the AWS backend purged the deletion task record faster than the cloud-nuke polling logic could check it.

In previous versions, if the poller received a NoSuchEntity response, it would treat the situation as an error. The updated logic in v0.52.0 is smarter. When it encounters this response, it performs a GetRole call to verify the actual state of the role. If the role is truly gone, the tool treats the deletion as a success. This change reduces noise and prevents unnecessary execution failures during cleanup runs.

Library updates for Go developers

Engineers using cloud-nuke as a library in their own Go automation tools need to be aware of a breaking change in the internal API. The functions aws.NukeAllResources and gcp.NukeAllResources have been updated to support the new concurrency features.

Both functions now require an additional parallelism int parameter. This allows library consumers to programmatically set the same concurrency limits that the CLI provides. While this requires a small code update for those importing these packages, it enables the same performance benefits in custom automation scripts.

Internal refactoring and consistency

Beyond the user facing features, the maintainers have performed internal cleanup to support the move toward a more concurrent architecture. The preservation of within region dependency order while allowing cross region parallelism shows a careful approach to maintaining the reliability of the tool while improving speed.

The update to the service linked role logic specifically targets a known sharp edge in the AWS API. By handling the transient nature of task records, the tool becomes more resilient to the eventual consistency quirks of the underlying cloud provider.

Where to get it

You can download the latest binaries and view the full changelog at the following locations:



denis256 at denis256.dev