Tharga.MongoDB 2.10.13

Tharga.MongoDB

A MongoDB repository toolkit for .NET 8 / 9 / 10. Adds dynamic database/collection naming, automatic index assurance, document-level locking with commit-on-success semantics, multi-document transactions, keyset pagination, and built-in monitoring of every call (latency, slow queries, cache stats) on top of the official MongoDB.Driver.

Install

dotnet add package Tharga.MongoDB

builder.AddMongoDB();

Configure connection strings in appsettings.json:

"ConnectionStrings": {
  "Default": "mongodb://localhost:27017/MyApp{Environment}{Part}"
}

Define an entity and a repository collection — both get auto-registered into DI:

public record Order : EntityBase<ObjectId>
{
    public string CustomerName { get; init; }
    public DateTime CreatedAt { get; init; }
}

public class OrderCollection : DiskRepositoryCollectionBase<Order, ObjectId>
{
    public OrderCollection(IMongoDbServiceFactory factory) : base(factory) { }
    public override string CollectionName => "orders";
}

Inject IRepositoryCollection<Order, ObjectId> (or your concrete collection) wherever you need it.

What's in the box

• Repository pattern with Disk and Lockable collection bases. Lockable adds per-document optimistic locks with commit/release/error workflows so you can hand a document to a worker, let it crash, and have the lock auto-recover.
• Dynamic database & collection naming via DatabaseContext — multi-tenant URLs like mongodb://.../{Environment}{Part} resolve at call time, not at startup.
• Automatic index assurance with rename-safe modes (ByName, BySchema, DropCreate).
• Multi-document transactions via WithTransactionAsync — session-aware writes on both Disk and Lockable; lockable leases support transactional commit so the document update and your business writes land atomically.
• Keyset pagination — GetPageAsync / GetPageProjectionAsync with opaque CursorToken. O(log N) per page regardless of depth, no skip penalty. CursorPager<TEntity, TKey> adapter for grids that emit (skip, pageSize).
• Monitoring — every call (with filter, sort, explain plan) is captured and exposed via IDatabaseMonitor for live dashboards. Slow-query detection, queue metrics, cache stats, and per-collection status.
• Result limiter — hard cap on returned document counts via ResultLimit config.
• Execute limiter — auto-sized per connection pool to keep long-running streams from oversubscribing the driver.
• Companion packages: Tharga.MongoDB.Blazor for the admin UI, Tharga.MongoDB.Mcp for MCP/AI integration, Tharga.MongoDB.Monitor.Client + .Server for centralised monitoring across multiple agents.

Documentation

Full docs, configuration reference, and worked examples for each feature: github.com/Tharga/MongoDB.