[QUESTION] Proper errors handling

[kpeu3i]

Versions

Sarama	Kafka	Go
v1.27.0	2+	1.14.7

Configuration

We use SAAS kafka installation https://www.confluent.io/

config := sarama.NewConfig()
config.Version = sarama.V2_0_0_0

config.Net.SASL.Enable = t.config.SASL.Enabled
config.Net.SASL.User = t.config.SASL.User
config.Net.SASL.Password = t.config.SASL.Password
config.Net.SASL.Mechanism = sarama.SASLMechanism(t.config.SASL.Mechanism)
config.Net.TLS.Enable = t.config.TLS.Enabled

config.Metadata.Retry.Max = 10
config.Metadata.Retry.Backoff = time.Second

config.Consumer.MaxWaitTime = 1000 * time.Millisecond
config.Consumer.Return.Errors = true
config.Consumer.Group.Rebalance.Retry.Backoff = 5 * time.Second
config.Consumer.Group.Rebalance.Retry.Max = 10
// Possible fix for "The provided member is not known in the current generation" error.
// See https://github.com/Shopify/sarama/issues/1310#issuecomment-483768308
config.Consumer.Group.Session.Timeout = 20 * time.Second
config.Consumer.Group.Heartbeat.Interval = 6 * time.Second
config.Consumer.MaxProcessingTime = 500 * time.Millisecond

Can somebody clarify how to properly handle errors received from the consumer errors channel? Currently, it panics when an error occurred in the channel and this leads to the restart of an application.

go func() {
	for {
		select {
		case err := <-consumer.Errors():
			if err != nil {
				panic(err)
			}
		case <-t.quit:
			return
		}
	}
}()

Is possible to skip errors and continue listening to messages?

go func() {
	for {
		select {
		case err := <-consumer.Errors():
			if err != nil {
				log.Println(err)
			}
		case <-t.quit:
			return
		}
	}
}()

Why I am asking is because we see a lot of errors such:

kafka: error while consuming topic_name/partition: read tcp IP_1->IP_2: i/o timeout
kafka: error while consuming topic_name/partition: write tcp IP_1->IP_2: i/o timeout

[github-actions]

Thank you for taking the time to raise this issue. However, it has not had any activity on it in the past 90 days and will be closed in 30 days if no updates occur.
Please check if the main branch has already resolved the issue since it was raised. If you believe the issue is still valid and you would like input from the maintainers then please comment to ask for it to be reviewed.