Skip to main content

Distributed Object Storage

Distributed Object Storage is an S3-inspired object storage service for storing versioned files, issuing presigned download URLs, and replicating objects asynchronously across storage nodes.

Key Capabilities

  • object upload by bucket and key
  • immutable object versions with latest-version reads
  • presigned download URLs
  • PostgreSQL-backed metadata
  • durable asynchronous replication jobs
  • strict bucket and object-key validation

When To Use It

Use Distributed Object Storage when you need:

  • a file/object API rather than block or database storage
  • temporary signed download access
  • versioned object writes
  • durable replication work that can survive worker restarts

Core Concepts

ConceptMeaning
Bucketnamespace for stored objects
Object keyobject identifier within a bucket
Versionimmutable uploaded revision of an object
Presigned URLtemporary signed download URL
Replication jobdurable background task for secondary copies

Current Scope

The current release provides the core upload, presign, download, versioning, and replication workflow. It is not yet fully S3-compatible and does not yet include multipart upload, API-key authentication, lifecycle policies, or production-grade observability.

Documentation

  1. Getting Started: run the service and complete the first object flow.
  2. User Guide: product behavior and common workflows.
  3. Examples: practical API examples.
  4. API Reference: endpoint contract.
  5. Security and Access: current controls and gaps.
  6. Data Protection and Durability: replication guarantees.
  7. Monitoring and Troubleshooting: operational checks.

Source of Truth