You don't need an Event Sourcing framework. Or do you?
summary
Argues that teams usually end up building reusable plumbing for production event sourcing and so a .NET Event Store DB / Postgres / SQL Server library is practical; shows an open source framework that wires serialization, subscriptions, checkpointing, metrics and distributed tracing, and exposes a Blazor time‑machine UI that reconstructs state from events (demo mostly works with one hiccup). Covers projection pain points — stale queries, snapshotting, long streams — and a partitioned projections approach that raises throughput (tenfold concurrency example) for systems ingesting millions of events/day.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:14Event sourcing struggles
- 3:22Misconceptions and diagrams
- 5:54Production day reality check
- 7:59Solving technical challenges
- 10:20Infrastructure abstraction
- 14:45Framework maintenance issues
- 18:22Framework vs libraries
- 22:48Why companies open source
- 25:53The 'Simultaneous' library
- 30:40Code implementation details
- 34:09State reconstruction
- 36:03Foundational principles
- 38:55Demo and CQRS
- 43:24The time machine feature
- 50:01Optimizing projections
related talks
57mEventuous Event Sourcing with .NET
44m.NET: Practical Event Sourcing with Marten
45mUsing modern .NET technologies to decompose a monolith into micro-services
45mBuilding a C# Application from Scratch with CleanArchitecture, CQRS and Event Sourcing
47m.NET Blub: Frameworks beyond Microsoft
01h 07mSteeltoe and the Open Source .NET Renaissance
07h 42mDay 2: Event-Driven Systems & GenAI with Aspire, F#, dotMemory, and More
01h 14mEffortless Distributed Systems With .NET Aspire