New OpenStack Swift scaler

[domminiks]

My team (@sidilabs) and I want to contribute to KEDA by proposing and also implementing a RedHat/OpenStack Swift scaler.

More info about our proposal:

• Scaler Source: Swift is the default object-store for OpenStack. It works much like AWS S3 or Azure Blob Storage.

  The OpenStack Object Store project, known as Swift, offers cloud storage software so that you can store and retrieve lots of data with a simple API. It's built for scale and optimized for durability, availability, and concurrency across the entire data set.

• How do you want to scale: The idea is to scale based on the object count in a certain container. As the object count (number of files, folders, ...) changes, KEDA scales up or down accordingly.

• Authentication: for authentication we must use Keystone (another OpenStack project). With Keystone, one can get an access token to be used for authentication/authorization. We've investigated two auth methods for getting a token and then querying the Swift API:

  • Simple username/password
  • Using Application Credentials

We understand that both auth methods should be implemented by the scaler.

[zroubalik]

Looking good! That would be a great contribution.

Just a minor nit, implementing e2e tests with the scaler would be really great: https://github.com/kedacore/keda/tree/main/tests

[domminiks]

Ok, I took a note about it, thanks!

Letting the community know about our progress soon 😄

[tomkerkhove]

This was implemented, closing