MongoDB storage

Document-database backend. Pairs naturally with applications that already use MongoDB for their business data.

Install

dotnet add package OpStream.Server.Storage.MongoDB

Setup

builder.Services
    .AddOpStream()
    .UseMongoDbStorage(options =>
    {
        options.ConnectionString = builder.Configuration.GetConnectionString("Mongo")!;
        options.DatabaseName     = "opstream";
        // Optional collection name overrides
        options.OpsCollectionName       = "ops";
        options.SnapshotsCollectionName = "snapshots";
    });

Collection shape

// ops
{
  _id:        ObjectId,
  documentId: "...",
  revision:   42,
  authorId:   "peer-7",
  timestamp:  ISODate("..."),
  payload:    BinData,
  engineType: "TextOp"
}
// Compound index: { documentId: 1, revision: 1 }

// snapshots
{
  _id:        "documentId",   // _id is the document id
  revision:   42,
  timestamp:  ISODate("..."),
  state:      BinData
}

Performance notes

• The unique compound index on (documentId, revision) is created on first use. It's also the natural sort key for StreamOpsAsync.
• For multi-region MongoDB Atlas clusters, pin OpStream's connection string to the primary read preference; reading from secondaries risks reading a stale op log between writes.

When to pick this

• Business data is already in MongoDB — operational expertise is reused.
• Schemaless ergonomics for ad-hoc payload shapes.
• You need ACID transactions across multiple documents in the same OpStream session — use a relational backend.