Stratara.Outbox.AzureServiceBus
4.5.0
Prefix Reserved
See the version list below for details.
dotnet add package Stratara.Outbox.AzureServiceBus --version 4.5.0
NuGet\Install-Package Stratara.Outbox.AzureServiceBus -Version 4.5.0
<PackageReference Include="Stratara.Outbox.AzureServiceBus" Version="4.5.0" />
<PackageVersion Include="Stratara.Outbox.AzureServiceBus" Version="4.5.0" />
<PackageReference Include="Stratara.Outbox.AzureServiceBus" />
paket add Stratara.Outbox.AzureServiceBus --version 4.5.0
#r "nuget: Stratara.Outbox.AzureServiceBus, 4.5.0"
#:package Stratara.Outbox.AzureServiceBus@4.5.0
#addin nuget:?package=Stratara.Outbox.AzureServiceBus&version=4.5.0
#tool nuget:?package=Stratara.Outbox.AzureServiceBus&version=4.5.0
Stratara.Outbox.AzureServiceBus
Derived. The behaviour described here is specified under
openspec/specs/. Those specifications are the source; this page explains and illustrates them.
License: MIT.
Azure Service Bus implementation of Stratara.Abstractions.Messaging.IMessageBus. Publishes JSON-serialized messages to topics and exposes a subscription helper that wires up a Service Bus processor with per-message exception classification:
- success →
CompleteMessageAsync ConcurrencyException→AbandonMessageAsyncuntilMessageRetry:MaxConflictRequeuesredeliveries (default 100) have not resolved it, thenDeadLetterMessageAsyncwith reasonconflict- any other exception →
AbandonMessageAsyncuntilMessageRetry:MaxDeliveryAttemptsdeliveries (default 3) have failed, thenDeadLetterMessageAsyncwith reasonfailureand the exception in the description
The subscription's own MaxDeliveryCount must leave room for the bounds; a subscription created with Service Bus defaults allows 10 deliveries. Where the host can read the subscription, the bus lowers the bounds for it to fit and logs 108_111.
System-level errors (connection drops, auth failures) arrive via ProcessErrorAsync and are logged; the Service Bus client owns the reconnect / retry policy for those.
A subscription that stops closes its processor, which takes no further message and waits for its running handler — as long as the host's shutdown timeout allows while the application stops, twenty seconds otherwise; the host waits for that before it counts as stopped. A handler that gives up because the subscription stops has its message abandoned rather than dead-lettered by the framework; the broker counts that delivery like any other. A handler's outcome is settled whatever the subscription's own cancellation says, and a handler that failed because its save committed but could not publish has its message completed, not delivered again.
Install
dotnet add package Stratara.Outbox.AzureServiceBus
Register the bus in your DI composition — one call wires the ServiceBusClient, the envelope
options and the IMessageBus implementation:
// Connection string:
builder.Services.AddAzureServiceBus(builder.Configuration.GetConnectionString("ServiceBus")!);
// Or, preferred in Azure — managed identity, no secret in configuration:
builder.Services.AddAzureServiceBusWithManagedIdentity("my-namespace.servicebus.windows.net");
The AzureServiceBusBus implementation is internal; register it through these extensions rather
than by naming the type.
The explicitly-chosen transport wins.
AddAzureServiceBusreplaces theIMessageBusregistration, so it takes effect even when the RabbitMQ umbrella (builder.AddMessaging(), which the worker composites call) already claimed the slot. Still use one transport per host — calling both is a configuration smell — but an explicitAddAzureServiceBuswill no longer be a silent no-op behind a composite.
Notes
Pre-3.0 this implementation lived inside Stratara.Outbox.RabbitMQ. As of 3.0 the two transports are separate packages so a consumer who only wants RabbitMQ does not drag the Azure Service Bus SDK into the dependency tree (and vice versa).
| 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
- Azure.Identity (>= 1.21.0)
- Azure.Messaging.ServiceBus (>= 7.21.0)
- Microsoft.Extensions.DependencyInjection.Abstractions (>= 10.0.13)
- Microsoft.Extensions.Hosting.Abstractions (>= 10.0.13)
- Microsoft.Extensions.Logging.Abstractions (>= 10.0.13)
- Microsoft.Extensions.Options (>= 10.0.13)
- Microsoft.Extensions.Options.ConfigurationExtensions (>= 10.0.13)
- Stratara.Abstractions (>= 4.5.0)
- Stratara.Shared (>= 4.5.0)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 4.8.0 | 103 | 11/6/2026 |
| 4.7.1 | 97 | 11/5/2026 |
| 4.7.0 | 106 | 10/30/2026 |
| 4.6.1 | 105 | 10/29/2026 |
| 4.6.0 | 102 | 10/28/2026 |
| 4.5.0 | 107 | 10/28/2026 |
| 4.4.10 | 103 | 10/26/2026 |
| 4.4.9 | 98 | 10/26/2026 |
| 4.4.8 | 100 | 10/24/2026 |
| 4.4.7 | 102 | 10/23/2026 |
| 4.4.6 | 97 | 10/22/2026 |
| 4.4.5 | 108 | 10/14/2026 |
| 4.4.4 | 107 | 10/10/2026 |
| 4.4.3 | 111 | 10/10/2026 |
| 4.4.2 | 113 | 10/2/2026 |
| 4.4.1 | 107 | 9/29/2026 |
| 4.4.0 | 109 | 9/27/2026 |
| 4.3.1 | 116 | 9/25/2026 |
| 4.3.0 | 113 | 9/23/2026 |
| 4.2.0 | 123 | 9/18/2026 |
A release about reads that cost more the more the store holds. Loading an aggregate read its stream and its
snapshot through every partition of the store, so every command grew slower as the store grew. A write context
that applies `ApplyGlobalEntityConfiguration()` also lost the index on the event stream's sequence number, which
left the hashing worker, reading in sequence order and the store's head without an index. Every read of one stream
is now confined to its partition, and the write model declares the indexes it was missing, whichever key the
context gives the event stream. No public type or member changes, but the write context needs a migration, and a
write context that overrides `ConfigureConventions` without calling `base` is now refused at start-up.
**Upgrading:**
- **If your store holds events written before 3.0.0, check it first.** Releases before 3.0.0 worked out a
stream's partition with another formula, and the store now reads a stream only in the partition its id gives.
The composition guide (*A store written before 3.0.0*) gives a check that needs nothing but the database, and a
repair that moves the rows to their partition without touching the hash chain. A store that 3.0.0 or later wrote
from its start has nothing to check.
- **Generate an EF Core migration for your write context.** It adds up to three indexes and changes no table or
column: `ix_event_stream_entry_sequence_number` (unique, only where the event stream is not keyed by its sequence
number, which `ApplyGlobalEntityConfiguration()` makes it), `ix_event_stream_entry_unhashed_sequence_number`
(the entries whose hash is still null) and `ix_event_chain_anchor_sequence_number`. A host that does not migrate
fails at `Migrate()` on pending model changes, as with any schema addition.
- **If you declared these indexes yourself, remove your declarations first.** Under the same names and
definitions, the migration then leaves them as they are. A `(stream_id, version)` index you added to bridge the
stream reads is no longer needed and can be dropped.
- **On a large store, create the indexes without blocking writes.** The schema section of the composition guide
shows a migration that creates them with `CREATE INDEX CONCURRENTLY`.
- **A context that overrides `ConfigureConventions` calls `base.ConfigureConventions`.** `WriteDbContext<TContext>`
declares the sequence-number indexes there. A write context whose override does not call it is now refused when
its model is built, at start-up and in `dotnet ef` alike, with an `InvalidOperationException` that names the
context, so it never runs or migrates without the indexes. An override that was fine on 4.4 because the
method was empty there needs `base.ConfigureConventions(configurationBuilder);` now.
- **An entry or snapshot written outside the framework has to carry its stream's partition.** The store writes
each one with the partition its stream id gives it, through `BucketCalculator.GetBucketId`, and now reads one
stream only there. A row a host inserted itself under another partition, in a test fixture or an import, is no
longer found by the reads of its stream.
### Fixed
- **Loading an aggregate costs what its stream holds.** The reads of one stream (whether it exists, its first entry,
a range of versions, its current version) and of its snapshot conditioned on the stream id alone. The indexes that
serve them begin with the partition, so PostgreSQL before 18 walked the whole index for each of them, and a
command loads several. They now condition on the partition as well. On 300,000 events, reading a stream of three
went from 2,404 pages to 6.
- **The sequence number stays indexed under `ApplyGlobalEntityConfiguration()`.** That convention keys every entity
by `Id`, the event stream included, which dropped the key the framework had put on the sequence number. The
hashing worker's reads, reading the store in sequence order and the store's head then scanned the whole table on
every pass. The write model now declares a unique index on the sequence number wherever it is not the key.
- **The hashing worker reads only what is still unchained.** An index over the entries whose hash is null lets
each pass read the backlog and not the chained history before it. Its filter names the column your model maps the
hash to.
- **The latest anchor is found by index.** The hash chain's anchors are indexed by sequence number, so finding the
last one no longer sorts every anchor on every pass.