Stratara.Outbox.RabbitMQ
4.4.1
Prefix Reserved
See the version list below for details.
dotnet add package Stratara.Outbox.RabbitMQ --version 4.4.1
NuGet\Install-Package Stratara.Outbox.RabbitMQ -Version 4.4.1
<PackageReference Include="Stratara.Outbox.RabbitMQ" Version="4.4.1" />
<PackageVersion Include="Stratara.Outbox.RabbitMQ" Version="4.4.1" />
<PackageReference Include="Stratara.Outbox.RabbitMQ" />
paket add Stratara.Outbox.RabbitMQ --version 4.4.1
#r "nuget: Stratara.Outbox.RabbitMQ, 4.4.1"
#:package Stratara.Outbox.RabbitMQ@4.4.1
#addin nuget:?package=Stratara.Outbox.RabbitMQ&version=4.4.1
#tool nuget:?package=Stratara.Outbox.RabbitMQ&version=4.4.1
Stratara.Outbox.RabbitMQ
Derived. The behaviour described here is specified under
openspec/specs/. Those specifications are the source; this page explains and illustrates them.
License: MIT.
Outbox-pattern command + event dispatch for the Stratara event-sourced stack with a RabbitMQ message-bus implementation. Contains the write-side dispatchers, the outbox-retry worker, the read-side mediator command worker, the RabbitMQ bus, and the ProjectionReplayState that coordinates dispatch skip during projection replay — shared over Redis where a connection is registered, held in process otherwise. Azure Service Bus ships separately as Stratara.Outbox.AzureServiceBus.
What's in the box
| Folder | Contents |
|---|---|
Outbox/ |
OutboxOptions, CommandOutboxDispatcher (write-side ICommand fan-out via IMessageBus, falls back to outbox table on bus failure), EventBundleOutboxDispatcher (same for EventBundle), OutboxWorker (hosted service that retries unpublished outbox rows on a polling interval), NullOutboxLock + RedisOutboxLock (IOutboxLock implementations — default no-op for single-instance deployments, Redis-leased distributed lock for multi-replica setups) |
Messaging/ |
RabbitMqBus — IMessageBus over RabbitMQ. A worker subscription is a quorum queue <subscription>.v2 with <subscription>.dead-letter beside it; a message a handler cannot take is redelivered under MessageRetryOptions and then dead-lettered. A subscription holds at most Messaging:PrefetchCount messages (default 16), and the host waits for a stopping subscription's running handler to settle before it counts as stopped — within the host's shutdown timeout; twenty seconds for a subscription stopped on its own. Azure Service Bus ships as the sibling Stratara.Outbox.AzureServiceBus package. |
Mediator/ |
MediatorCommandWorker (hosted service that subscribes to the command topic and dispatches into the in-process IMediator) |
Projections/ |
ProjectionReplayState (concrete IProjectionReplayState, Redis-backed where a connection is registered, in process otherwise; dispatchers skip publishing while replay is active), ProjectionReplayOptions (how long the replay marking survives without renewal) |
DependencyInjection/ |
AddOutboxDispatcher(), AddOutboxWorker(IConfiguration), AddRedisOutboxLock() (opt-in distributed lock), AddProjectionReplayState(), AddMediatorWorker(), AddMessaging() |
Diagnostics/Extensions/ |
LoggerOutboxExtensions, LoggerMessagingExtensions (source-generated logger surfaces) |
Quick start
// In your API host:
builder.AddMessaging(); // IMessageBus + MessagingOptions binding
builder.Services
.AddOutboxDispatcher() // CommandOutboxDispatcher + EventBundleOutboxDispatcher + ProjectionReplayState
.AddOutboxWorker(builder.Configuration); // OutboxWorker hosted service (only if this host owns retries)
// In your command worker:
builder.Services
.AddMediatorWorker(); // MediatorCommandWorker hosted service
The dispatchers consult IProjectionReplayState.IsReplayActive before each publish and skip dispatch (writing to the outbox table only) while a replay is in progress. They also record the outbox.published counter for what the broker accepted — the worker counts nothing, because only the dispatcher knows what went out.
The replay marking and its progress counters are held on the lease configured by ProjectionReplayOptions.LeaseSeconds (default 300), which the replay renews every time it reports progress. A replay whose host is killed stops renewing it and the marking lapses on its own, instead of suppressing publication indefinitely. AddProjectionReplayState() reads the options from the ProjectionReplay section of the host's configuration; services.Configure<ProjectionReplayOptions>(...) after that call takes precedence. A lease of zero or less fails the host at start. Set it longer than the slowest stretch between two progress reports — too short lets the marking lapse while the replay is still running.
A drain cycle takes one batch of each kind and ends. Rows the broker would not accept stay in the table for the next interval; a cycle never re-reads what it just failed to publish, so an unreachable broker or a suppressed drain cannot turn a cycle into a loop over the same rows.
Multi-instance outbox workers
AddOutboxWorker registers NullOutboxLock as the default IOutboxLock — a no-op that preserves the single-instance assumption. For multi-replica deployments call AddRedisOutboxLock() afterwards; it replaces the no-op with a Redis-leased lock (SET stratara:outbox:lock NX EX) so only one replica drains at a time:
builder.AddCaching(); // registers IConnectionMultiplexer
builder.Services
.AddOutboxDispatcher()
.AddOutboxWorker(builder.Configuration)
.AddRedisOutboxLock(); // multi-replica safe
The lease defaults to 60 s (OutboxOptions.LockLeaseSeconds). Tune it so it exceeds the worst-case drain duration; otherwise the lock can expire mid-cycle and a peer may start a concurrent drain. Outbox semantics are still at-least-once, so a duplicate publish is recoverable provided handlers stay idempotent.
Dependencies
Stratara.Abstractions— forICommand,IEvent,IMessageBus,ICommandOutboxDispatcher,IEventBundleOutboxDispatcher,IProjectionReplayState,IMessagingIdentifier,IWriteUnitOfWork(used at runtime via the outbox repository).Stratara.Contracts— forEventBundle+CommandEnvelopemessages.Stratara.Mediator—MediatorCommandWorkerdispatches into the in-processIMediator.Stratara.Sessions— dispatcher hydratesCommandEnvelopefrom the current session context.Stratara.Shared— for messaging primitives, resilience pipeline names, mapping helpers, and the diagnostics base.RabbitMQ.Client,StackExchange.Redis(only used when a connection is registered: shared replay state, optional outbox lock).Microsoft.Extensions.Hosting.Abstractions+Microsoft.Extensions.Options.ConfigurationExtensions— for hosted services + options binding.
The outbox dispatcher persists rows through
IWriteUnitOfWork.CreateOutboxRepository— that interface lives inStratara.Abstractions, but the concrete implementation comes fromStratara.EventSourcing.EntityFrameworkCore. Reference that package alongside this one to get a working stack.
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net10.0
- Microsoft.Extensions.Configuration.Abstractions (>= 10.0.11)
- Microsoft.Extensions.DependencyInjection (>= 10.0.11)
- Microsoft.Extensions.Hosting.Abstractions (>= 10.0.11)
- Microsoft.Extensions.Logging.Abstractions (>= 10.0.11)
- Microsoft.Extensions.Options.ConfigurationExtensions (>= 10.0.11)
- RabbitMQ.Client (>= 7.2.2)
- StackExchange.Redis (>= 3.1.31)
- Stratara.Abstractions (>= 4.4.1)
- Stratara.Contracts (>= 4.4.1)
- Stratara.Mediator (>= 4.4.1)
- Stratara.Sessions (>= 4.4.1)
- Stratara.Shared (>= 4.4.1)
NuGet packages (2)
Showing the top 2 NuGet packages that depend on Stratara.Outbox.RabbitMQ:
| Package | Downloads |
|---|---|
|
Stratara.Infrastructure
Infrastructure glue for the Stratara framework — authorization decorators, configuration providers, and DI composition helpers that wire Mediator, Outbox, Identity, and EF Core into a hosted app. |
|
|
Stratara.EventSourcing.WorkerDefaults
Worker-host wiring composites for the Stratara event-sourced stack. IHostApplicationBuilder extensions (AddBackendServices, AddCommandWorkerServices, AddHeavyCommandWorkerServices, AddEventProjectionWorkerServices, AddEventStreamHashWorkerServices, AddSagaWorkerServices, AddOutboxWorkerServices, and AddCommandServices, AddEventProjectionServices and AddSagaServices without the bus-fed workers) bundle the per-concern DI calls so each worker host opts in with one line. |
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 4.4.2 | 47 | 10/2/2026 |
| 4.4.1 | 334 | 9/29/2026 |
| 4.4.0 | 150 | 9/27/2026 |
| 4.3.1 | 142 | 9/25/2026 |
| 4.3.0 | 313 | 9/23/2026 |
| 4.2.0 | 182 | 9/18/2026 |
| 4.1.1 | 110 | 9/16/2026 |
| 4.1.0 | 115 | 9/16/2026 |
| 4.0.4 | 946 | 9/14/2026 |
| 4.0.3 | 191 | 9/3/2026 |
| 4.0.2 | 615 | 9/3/2026 |
| 4.0.1 | 295 | 9/2/2026 |
| 4.0.0 | 175 | 8/31/2026 |
| 4.0.0-preview.1 | 83 | 8/31/2026 |
| 3.4.0 | 164 | 8/28/2026 |
| 3.3.0 | 133 | 8/25/2026 |
| 3.2.3 | 146 | 8/22/2026 |
| 3.2.2 | 148 | 8/14/2026 |
| 3.2.1 | 159 | 8/2/2026 |
| 3.2.0 | 148 | 7/18/2026 |
A release about errors that were not errors. Where two writers insert the same row at once — two
deliveries forgetting a deleted tenant, two activations writing a partition's first checkpoint, two
hosts importing one API key — the framework already recovered, but EF Core had logged the losing
statement at Error first; on PostgreSQL and SQLite the insert now skips a row that is already there.
And a RabbitMQ bus whose broker went away first no longer reports a warning for every subscription it
stops, nor a failed handler for a message it merely could not acknowledge. No migration and no
configuration change is needed.
### Fixed
- **Forgetting a deleted tenant leaves no error in the log when two deliveries race.** A tenant's two
deletion facts arrive in two bundles, so two deliveries of one projection can record the same tenant
at once. The losing insert was caught and the tenant recorded all the same, but EF Core had already
logged the failed statement at Error (`20102` and `10000`), so a customer deletion could leave two
errors in a projection worker's log. On PostgreSQL and SQLite the tenants are now recorded with an
insert that skips a row already there; on other providers the store still retries, and EF Core still
logs the statement that lost.
- **A projection checkpoint's first write, and an API key two hosts import at once, leave no error in
the log.** Two activations of one reader that overlap during a failover can both write a partition's
first checkpoint, and hosts that seed the same key as they boot import it at once. The writer whose
insert lost recovered — it took the other's checkpoint over, or adopted the other's key — but EF Core
had already logged the failed statement at Error. On PostgreSQL and SQLite the insert now skips a row
that is already there, as the record of forgotten tenants does; a checkpoint that another writer
inserted and removed again before it could be taken over is refused with an
`InvalidOperationException`. An import takes that path only where it is the only write on its
context, and then writes the key and its membership in one transaction of its own, outside
`SaveChanges` and its interceptors. An import on a context with unsaved changes or inside a
transaction of the caller's saves through `SaveChanges` as before, and so does a table a consumer
widened with columns the row does not carry, and every other provider: there the writer still
recovers, and EF Core still logs the statement that lost.
- **A RabbitMQ subscription whose broker is already gone stops without a warning.** When a stack stops
its broker before its hosts, a stopping subscription's cancel met a connection that had already
closed, or that closed while the cancel waited for its reply, and every subscription of every host
logged `108_107` at Warning — new in 4.4.0, which made a subscription cancel its consumer before it
closes. A closed connection has no consumer left to cancel, which is where the stop was going, so
this is now logged at Debug as `108_007` (`LogEvents.Messaging.SubscriptionAlreadyClosed`). A cancel
that fails while the connection is still up still warns.
- **A handler that completes after the broker went away is no longer reported as failed.** Its
acknowledgement met a closed channel, and the RabbitMQ bus logged `108_101` at Error although the
handler had succeeded, then tried to reject the message on the same closed channel. It now logs
`108_008` (`LogEvents.Messaging.MessageNotSettledChannelClosed`) at Information: the message could
not be settled, and the broker delivers it again.