nap_all_day

fix: race condition(may panic) when closing consumer group

2

# Background We encounter a data race issue when closing consumer group. ##### Versions *Please specify real version numbers or git SHAs, not just "Latest" since that changes fairly regularly.*...

As shown below, Kafka supports specifying multiple rebalance protocol from the very begining. ([source](https://cwiki.apache.org/confluence/display/KAFKA/A+Guide+To+The+Kafka+Protocol#AGuideToTheKafkaProtocol-PossibleErrorCodes:~:text=all%20client%20implementations.-,Join%20Group%20Request,-The%20join%20group)) ``` v0 supported in 0.9.0.0 and greater JoinGroupRequest => GroupId SessionTimeout MemberId ProtocolType GroupProtocols GroupId =>...

feat(consumer): incremental cooperative balance strategy (KIP-429)

21

## Incremental Cooperative Protocol POC Proposal I'm excited to present this Proof of Concept (POC) version that supports the incremental cooperative protocol as outlined in [KIP-429](https://cwiki.apache.org/confluence/display/KAFKA/KIP-429%3A+Kafka+Consumer+Incremental+Rebalance+Protocol). The proposed feature addresses...

fix #2256 violates the functionality of `config.Metadata.Retry.Max`

4

##### Versions *Please specify real version numbers or git SHAs, not just "Latest" since that changes fairly regularly.* | Sarama | Kafka | Go | |--------|-------|----| | v1.40.1 | N/A...

remove: redundant backoff timer in heartbeat loop

4

The timer `retryBackoff` here is redundant. Because in the underlying logic of `func Coordinator(consumerGroup string) (*Broker, error))`, `findCoordinator` has already utilized a retry mechanism with `client.conf.Metadata.Retry.Backoff` or `client.conf.Metadata.Retry.BackoffFunc`. ``` func...

`arbiter` will tick periodically forever. Normally it wouldn't cause any issue. But it's very inconvenient for users if Goroutine leak detector(like https://github.com/uber-go/goleak) is used in testing phase. btw, I also...