Releases: rabbitmq/rabbitmq-stream-dotnet-client
v1.8.4
Bug Fixes
- Use user defined loopback address if available (#374) by @CoenraadS in #377
New Contributors
- @CoenraadS made their first contribution in #377
Full Changelog: v1.8.3...v1.8.4
v1.8.3
Bug Fixes
Full Changelog: v1.8.2...v1.8.3
v1.8.2
Enhancements
- Add support build for net8 by @Gsantomaggio in #366
- Add missing ReadUuid. by @bangjiehan in #367
Bug Fixes
- Updated Secret try to reconnect the client by @Gsantomaggio in #364
- Read VersionString from AssemblyInformationalVersionAttribute (#365) by @doerig in #368
New Contributors
Thank you to:
- @bangjiehan made their first contribution in #367
- @doerig made their first contribution in #368
And to @vivek-singh2 for testing the client
Full Changelog: v1.8.1...v1.8.2
v1.8.1
Bug fix
- Dispatch the update secrets by @Gsantomaggio in #362
Full Changelog: v1.8.0...v1.8.1
This version is a bug fix for the 1.8.0
Please read the full release notes here:
https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/releases/tag/v1.8.0
v1.8.0
(Below, you can find all the changes/PRs)
What's new in 1.8
The 1.8 focus are:
- Multiple Consumers and Producers per connection
- Improve the reconnection for stream and super stream.
The high-level classes Consumer
and Producer
don't introduce breaking changes.
The RawSuperStream*
classes change the default behaviour. Please take a look at the section 3.
Main changes:
1. Multiple Consumers and Producers per connection
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig {
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 200 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 200 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 200
.
2. Improve the reconnections
Handle streamNotAvailable
, Add disconnection Info: #343
Improve the super stream reconnection: #344
Increase the backoff strategy time: #345
Please follow this document If you want to know more about what happens during a broker restart.
The focus is to improve the reconnection during the cluster restart.
3. Raw Super stream events
Removed the auto-reconnect. The RawSuperStreamProducer
and RawSuperStreamConsumer
classes now expose two events:
**NOTE: If you are using these classes, the auto-reconnect is removed to be compliant with all the Raw*
classes. **
You should use Consumer
and Producer
unless for a specific use case.
For Raw* users:
- Super Stream: during the disconnection, it is possible to understand the disconnection cause and reconnect the stream like:
var consumer = await system.CreateSuperStreamConsumer(configuration);
var completed = new TaskCompletionSource<bool>();
configuration.ConnectionClosedHandler = async (reason, stream) =>
{
if (reason == ConnectionClosedReason.Unexpected)
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false)
);
completed.SetResult(true);
}
};
The same is true for the standard consumer.
- Metadata update
MetadataHandler = async update =>
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false));
}
4. Add events to Producer
and Consumer
classes
See: #349
See also: https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/tree/main/docs/ReliableClient
where you can find an example of how to use StatusChanged
producerConfig.StatusChanged += (status) =>
{
var streamInfo = status.Partition is not null
? $" Partition {status.Partition} of super stream: {status.Stream}"
: $"Stream: {status.Stream}";
lp.LogInformation("Producer: {Id} - status changed from {From} to {To}. {Info}",
status.Identifier,
status.From,
status.To, streamInfo);
if (status.To == ReliableEntityStatus.Open)
{
publishEvent.Set();
}
else
{
publishEvent.Reset();
}};
5. Update Secret
See #342
6. SuperStream Creation/Deletion
See: #357
There are two ways to create the super-stream:
- Based on partitions with:
const string SuperStream = "my-first-system-super-stream";
var spec = new PartitionsSuperStreamSpec(SuperStream, 2);
await system.CreateSuperStream(spec);
- Based on keys with:
const string SuperStream = "countries";
var system = await StreamSystem.Create(new StreamSystemConfig());
var conf = new BindingsSuperStreamSpec(SuperStream, new[] { "italy", "france", "spain", "germany", "uk" });
await system.CreateSuperStream(conf);
Enhancements
- Multiple Producers and Consumers per connection by @Gsantomaggio in #328
- New event to handle Metadata update by @Gsantomaggio in #332
- Improve the rawProducer and rawSuperStreamProducer status by @Gsantomaggio in #337
- Handle cancellation token during the consumer close by @Gsantomaggio in #339
- Update license to Broadcom by @Gsantomaggio in #341
- Improve Reconnection by @Gsantomaggio in #343
- Improve the super stream reconnection by @Gsantomaggio in #344
- Increase the backoff strategy time by @Gsantomaggio in #345
- Add Producer/Consumer Identifier by @Gsantomaggio in #348
- Add event status changed by @Gsantomaggio in #349
- Update secret functionality by @simone-fariselli in #342
- Consumer connected only to the followers by @Gsantomaggio in #352
- Check the entity status during the close by @Gsantomaggio in #353
- Add Chunk info consumer side by @Gsantomaggio in #355
- Remove code duplication to handle the pool by @Gsantomaggio in #356
- Implement Add/Remove super stream feature by @Gsantomaggio in #357
- Add super stream exists API by @Gsantomaggio in #358
Bug fix
New Contributors
Full Changelog: v1.7.4...v1.8.0
v1.8.0-rc.3
Enhancements
- Consumer connected only to the followers by @Gsantomaggio in #352
- Check the entity status during the close by @Gsantomaggio in #353
- Add Chunk info consumer side by @Gsantomaggio in #355
- Remove code duplication to handle the pool by @Gsantomaggio in #356
Full Changelog: v1.8.0-rc.2...v1.8.0-rc.3
What's new in 1.8
The 1.8 focus are:
- Multiple Consumers and Producers per connection
- Improve the reconnection for stream and super stream.
The high-level classes Consumer
and Producer
don't introduce breaking changes.
The RawSuperStream*
classes change the default behaviour. Please take a look at the section 3.
1. Multiple Consumers and Producers per connection
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig {
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 200 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 200 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 200
.
2. Improve the reconnections
Handle streamNotAvailable
, Add disconnection Info: #343
Improve the super stream reconnection: #344
Increase the backoff strategy time: #345
Please follow this document If you want to know more about what happens during a broker restart.
The focus is to improve the reconnection during the cluster restart.
3. Raw Super stream events
Removed the auto-reconnect. The RawSuperStreamProducer
and RawSuperStreamConsumer
classes now expose two events:
**NOTE: If you are using these classes, the auto-reconnect is removed to be compliant with all the Raw*
classes. **
You should use Consumer
and Producer
unless for a specific use case.
For Raw* users:
- Super Stream: during the disconnection, it is possible to understand the disconnection cause and reconnect the stream like:
var consumer = await system.CreateSuperStreamConsumer(configuration);
var completed = new TaskCompletionSource<bool>();
configuration.ConnectionClosedHandler = async (reason, stream) =>
{
if (reason == ConnectionClosedReason.Unexpected)
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false)
);
completed.SetResult(true);
}
};
The same is true for the standard consumer.
- Metadata update
MetadataHandler = async update =>
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false));
}
4. Add events to Producer
and Consumer
classes
See: #349
See also: https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/tree/main/docs/ReliableClient
where you can find an example of how to use StatusChanged
producerConfig.StatusChanged += (status) =>
{
var streamInfo = status.Partition is not null
? $" Partition {status.Partition} of super stream: {status.Stream}"
: $"Stream: {status.Stream}";
lp.LogInformation("Producer: {Id} - status changed from {From} to {To}. {Info}",
status.Identifier,
status.From,
status.To, streamInfo);
if (status.To == ReliableEntityStatus.Open)
{
publishEvent.Set();
}
else
{
publishEvent.Reset();
}};
5. Update Secret
See #342
v1.8.0-rc.2
Enhancements
- Update secret functionality by @simone-fariselli in #342
Full Changelog: v1.8.0-rc.1...v1.8.0-rc.2
What's new in 1.8
The 1.8 focus are:
- Multiple Consumers and Producers per connection
- Improve the reconnection for stream and super stream.
The high-level classes Consumer
and Producer
don't introduce breaking changes.
The RawSuperStream*
classes change the default behaviour. Please take a look at the section 3.
1. Multiple Consumers and Producers per connection
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig {
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 200 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 200 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 200
.
2. Improve the reconnections
Handle streamNotAvailable
, Add disconnection Info: #343
Improve the super stream reconnection: #344
Increase the backoff strategy time: #345
Please follow this document If you want to know more about what happens during a broker restart.
The focus is to improve the reconnection during the cluster restart.
3. Raw Super stream events
Removed the auto-reconnect. The RawSuperStreamProducer
and RawSuperStreamConsumer
classes now expose two events:
**NOTE: If you are using these classes, the auto-reconnect is removed to be compliant with all the Raw*
classes. **
You should use Consumer
and Producer
unless for a specific use case.
For Raw* users:
- Super Stream: during the disconnection, it is possible to understand the disconnection cause and reconnect the stream like:
var consumer = await system.CreateSuperStreamConsumer(configuration);
var completed = new TaskCompletionSource<bool>();
configuration.ConnectionClosedHandler = async (reason, stream) =>
{
if (reason == ConnectionClosedReason.Unexpected)
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false)
);
completed.SetResult(true);
}
};
The same is true for the standard consumer.
- Metadata update
MetadataHandler = async update =>
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false));
}
4. Add events to Producer
and Consumer
classes
See: #349
See also: https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/tree/main/docs/ReliableClient
where you can find an example of how to use StatusChanged
producerConfig.StatusChanged += (status) =>
{
var streamInfo = status.Partition is not null
? $" Partition {status.Partition} of super stream: {status.Stream}"
: $"Stream: {status.Stream}";
lp.LogInformation("Producer: {Id} - status changed from {From} to {To}. {Info}",
status.Identifier,
status.From,
status.To, streamInfo);
if (status.To == ReliableEntityStatus.Open)
{
publishEvent.Set();
}
else
{
publishEvent.Reset();
}};
5. Update Secret
See #342
v1.8.0-rc.1
Enhancements
- Add Producer/Consumer Identifier by @Gsantomaggio in #348
- Add event status changed by @Gsantomaggio in #349
Full Changelog: v1.8.0-beta.2...v1.8.0-rc.1
What's new in 1.8
The 1.8 focus are:
- Multiple Consumers and Producers per connection
- Improve the reconnection for stream and super stream.
The high-level classes Consumer
and Producer
don't introduce breaking changes.
The RawSuperStream*
classes change the default behaviour. Please take a look at the section 3.
1. Multiple Consumers and Producers per connection
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig {
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 200 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 200 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 200
.
2. Improve the reconnections
Handle streamNotAvailable
, Add disconnection Info: #343
Improve the super stream reconnection: #344
Increase the backoff strategy time: #345
Please follow this document If you want to know more about what happens during a broker restart.
The focus is to improve the reconnection during the cluster restart.
3. Raw Super stream events
Removed the auto-reconnect. The RawSuperStreamProducer
and RawSuperStreamConsumer
classes now expose two events:
**NOTE: If you are using these classes, the auto-reconnect is removed to be compliant with all the Raw*
classes. **
You should use Consumer
and Producer
unless for a specific use case.
For Raw* users:
- Super Stream: during the disconnection, it is possible to understand the disconnection cause and reconnect the stream like:
var consumer = await system.CreateSuperStreamConsumer(configuration);
var completed = new TaskCompletionSource<bool>();
configuration.ConnectionClosedHandler = async (reason, stream) =>
{
if (reason == ConnectionClosedReason.Unexpected)
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false)
);
completed.SetResult(true);
}
};
The same is true for the standard consumer.
- Metadata update
MetadataHandler = async update =>
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false));
}
4. Add events to Producer
and Consumer
classes
See: #349
See also: https://github.com/rabbitmq/rabbitmq-stream-dotnet-client/tree/main/docs/ReliableClient
where you can find an example of how to use StatusChanged
producerConfig.StatusChanged += (status) =>
{
var streamInfo = status.Partition is not null
? $" Partition {status.Partition} of super stream: {status.Stream}"
: $"Stream: {status.Stream}";
lp.LogInformation("Producer: {Id} - status changed from {From} to {To}. {Info}",
status.Identifier,
status.From,
status.To, streamInfo);
if (status.To == ReliableEntityStatus.Open)
{
publishEvent.Set();
}
else
{
publishEvent.Reset();
}};
v.1.8.0-beta.2
Enhancements
- Improve the rawProducer and rawSuperStreamProducer status by @Gsantomaggio in #337
- Handle cancellation token during the consumer close by @Gsantomaggio in #339
- Update license to Broadcom by @Gsantomaggio in #341
- Improve Reconnection by @Gsantomaggio in #343
- Improve the super stream reconnection by @Gsantomaggio in #344
- Increase the backoff strategy time by @Gsantomaggio in #345
Full Changelog: v1.8.0-beta.1...v1.8.0-beta.2
What's new in 1.8
The 1.8 focus are:
- Multiple Consumers and Producers per connection
- Improve the reconnection for stream and super stream.
The high-level classes Consumer
and Producer
don't introduce breaking changes.
The RawSuperStream*
classes change the default behaviour. Please take a look at the section 3.
1. Multiple Consumers and Producers per connection
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig {
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 200 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 200 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig{
ConnectionPoolConfig = new ConnectionPoolConfig() {
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 200
.
2. Improve the reconnections
Handle streamNotAvailable
, Add disconnection Info: #343
Improve the super stream reconnection: #344
Increase the backoff strategy time: #345
Please follow this document If you want to know more about what happens during a broker restart.
The focus is to improve the reconnection during the cluster restart.
3. Raw Super stream events
Removed the auto-reconnect. The RawSuperStreamProducer
and RawSuperStreamConsumer
classes now expose two events:
**NOTE: If you are using these classes, the auto-reconnect is removed to be compliant with all the Raw*
classes. **
You should use Consumer
and Producer
unless for a specific use case.
For Raw* users:
- Super Stream: during the disconnection, it is possible to understand the disconnection cause and reconnect the stream like:
var consumer = await system.CreateSuperStreamConsumer(configuration);
var completed = new TaskCompletionSource<bool>();
configuration.ConnectionClosedHandler = async (reason, stream) =>
{
if (reason == ConnectionClosedReason.Unexpected)
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false)
);
completed.SetResult(true);
}
};
The same is true for the standard consumer.
- Metadata update
MetadataHandler = async update =>
{
await consumer.ReconnectPartition(
await system.StreamInfo(stream).ConfigureAwait(false));
}
v.1.8.0-beta.1
Enhancements
- Multiple Producers and Consumers per connection by @Gsantomaggio in #328
- New event to handle Metadata update by @Gsantomaggio in #332
Bug Fixes
To Read
The RabbitMQ stream protocol supports multi-producers and multi-consumers per TCP Connection.
This version introduces the connection pool for Consumers and Producers.
There is a new ConnectionPoolConfig
setting:
new StreamSystemConfig
{
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 10,
ProducersPerConnection = 10,
}
};
ConsumersPerConnection
== The number of consumers per connection min 1 max 255 default is 1
ProducersPerConnection
== The number of producers per connection min 1 max 255 default is 1
Each connection can handle different streams; see the image:
Performances
Sharing the same connection for multiple streams reduces the number of connections, but it could impact the performances:
- Consumer side. If one consumer is slow, it can also affect the other consumers
- Producer side: If all the producers are at full-rate, it can reduce the performances
The proper parameter depends on your environment.
Tip
You can use different StreamSystemConfig
like:
configToReduceTheConnections = new StreamSystemConfig
{
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 50, // high value
ProducersPerConnection = 50, // high value
}
}
configToIncreaseThePerformances = new StreamSystemConfig
{
ConnectionPoolConfig = new ConnectionPoolConfig()
{
ConsumersPerConnection = 1, // low value
ProducersPerConnection = 1, // low value
}
}
There are many combinations from 1
to 255
.
New Contributors
Full Changelog: v1.7.4...v1.8.0-beta.1