Add Consumers

In this section, we will learn how to add and configure a Consumer on KafkaFlow.

To add a consumer, you need to configure the Topic that the consumer will be listening to and the Consumer Group that will be part of. You can find an example below.

using KafkaFlow;
using Microsoft.Extensions.DependencyInjection;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .Topic("topic-name")
            .WithGroupId("sample-group")
            .WithBufferSize(100)
            .WithWorkersCount(10)
            .AddMiddlewares(middlewares => middlewares
                ...
            )
        )
    )
);

On a Consumer, you can configure the Middlewares that will be invoked. You can find more information on how to configure Middlewares here.

tip

You can use a naming pattern such as a wildcard to connect to any topic that matches a naming convention. You can find a sample on here.

Automatic Partitions Assignment

Using the Topic() or Topics() methods, the consumer will trigger the automatic partition assignment that will distribute the topic partitions across the application instances.

using KafkaFlow;
using KafkaFlow.Serializer;
using Microsoft.Extensions.DependencyInjection;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .Topic("topic-name")
            .WithGroupId("sample-group")
            ...
        )
    )
);

Manual Partitions Assignment

The client application can specify the topic partitions manually using the ManualAssignPartitions() method instead of using the Topic() or Topics() methods.

using KafkaFlow;
using KafkaFlow.Serializer;
using Microsoft.Extensions.DependencyInjection;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .ManualAssignPartitions("topic-name", new[] { 1, 2, 3 })
            ...
        )
    )
);

Offset Strategy

You can configure the Offset Strategy for a consumer group in case the Consumer Group has no offset stored in Kafka. There are two options:

• Earliest: Reads from the beginning of the Topic.
• Latest: Only reads new messages.

using KafkaFlow;
using Microsoft.Extensions.DependencyInjection;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .Topic("topic-name")
            .WithGroupId("sample-group")
            .WithAutoOffsetReset(AutoOffsetReset.Earliest)
            ...
        )
    )
);

Store Offsets Manually

By default, offsets are stored after the handler and middleware execution automatically.

To control Offset storing, it's possible to configure using WithManualMessageCompletion() as the following example:

using KafkaFlow;
using Microsoft.Extensions.DependencyInjection;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .Topic("topic-name")
            .WithGroupId("sample-group")
            .WithManualMessageCompletion()
            ...
        )
    )
);

Disable Store Offsets

In a scenario where storing offsets it's not needed, you can disable it by calling the WithoutStoringOffsets() method in the consumer setup.

using KafkaFlow;
using KafkaFlow.Serializer;
using Microsoft.Extensions.DependencyInjection;
using KafkaFlow.TypedHandler;

services.AddKafka(kafka => kafka
    .AddCluster(cluster => cluster
        .WithBrokers(new[] { "localhost:9092" })
        .AddConsumer(consumer => consumer
            .Topic("topic-name")
            .WithGroupId("sample-group")
            .WithoutStoringOffsets()
            ...
        )
    )
);