public async Task <IActionResult> ProduceMessage(CreateMessageRequest request)
{
if (!ModelState.IsValid)
{
return(BadRequest(ModelState));
}
await _service.ProduceAsync(request);
return(Ok());
}
protected override async Task ExecuteAsync(CancellationToken cancellationToken)
{
// You can find a lot of similar code on Confluent's examples.
// https://github.com/confluentinc/confluent-kafka-dotnet/blob/master/examples/Consumer/Program.cs
using var consumer = _kafkaService.GetConsumer();
try
{
consumer.Subscribe(topics);
while (true)
{
try
{
// setting the time span to 5 seconds so this will exit if there is no data after 5 seconds.
// this happens during the initial start.
var consumeResult = consumer.Consume(TimeSpan.FromSeconds(5));
if (consumeResult == null)
{
break;
}
if (consumeResult.IsPartitionEOF)
{
Console.WriteLine($"Reached end of topic {consumeResult.Topic}, partition {consumeResult.Partition}, offset {consumeResult.Offset}.");
break;
}
Console.WriteLine($"Received message at {consumeResult.TopicPartitionOffset}: {consumeResult.Message.Value}");
// setting
var setting = JsonSerializer.Deserialize <ProcessSettings>(consumeResult.Message.Value);
setting.LastProcessed = consumeResult.Message.Timestamp.UtcDateTime;
processSettings.Add(setting);
}
catch (ConsumeException e)
{
Console.WriteLine($"Consume error: {e.Error.Reason}");
}
}
}
catch (OperationCanceledException)
{
Console.WriteLine("Closing consumer.");
consumer.Close();
}
while (!cancellationToken.IsCancellationRequested)
{
// update setting
ProcessSettings setting = new ProcessSettings();
setting.SettingName = "KafkaLogCompaction";
setting.Value = "value";
var json = JsonSerializer.Serialize(setting);
await _kafkaService.ProduceAsync(topics, json, cancellationToken);
_logger.LogInformation("Worker running at: {time}", setting.LastProcessed);
await Task.Delay(2000, cancellationToken);
}
}
