Replace minio with alternatives

[simonswine]

While our helm chart uses a fairly old version of minio anyhow, we likely want to move on from minio, given they stopped providing helm charts and docker images for the OSS community.

Likely alternatives we should verify are:

• https://git.deuxfleurs.fr/Deuxfleurs/garage
• https://github.com/versity/versitygw/
• https://github.com/seaweedfs/seaweedfs

[ArshiyaDev]

Hi,

I'm interested in working on this issue. Before I start, I'd like to clarify the scope:

1. Are you looking for a full implementation, or would an initial research phase (comparing the three alternatives with recommendations) be valuable?

2. If broken into phases, I could handle:

   • Comparative analysis of Garage, VersityGW, and SeaweedFS
   • Testing/validation with the existing setup
   • Implementation/migration
   • Documentation

Which approach would be most helpful? Also, is there a timeline or priority level for this work?

Thanks.

[simonswine]

There is no timeline or deadline for this.

I think a comparison could be quite helpful before diving into more work.

The scope of this ticket is to replace the current usage of the minio helm chart in our pyroscope helm chart and replace it with an alternative: https://github.com/grafana/pyroscope/blob/2bc74cc802b15da57c4ad8e187f8ce05254a8bb0/operations/pyroscope/helm/pyroscope/Chart.yaml#L19-L23

While the implementation itself doesn't need to be production ready, it would be helpful if the solution we choose can support larger scale and a production grade service.

[ArshiyaDev]

I'd like to take this on.

Understanding that a comparison would be valuable before implementation, my plan is:

Research Phase: Evaluate Garage, SeaweedFS, and VersityGW against key criteria:

• Production-grade scalability
• Helm chart availability and maintenance
• S3 API compatibility
• Community activity and support
• Resource overhead

Could you assign this issue to me?