Struct Consumer 

pub struct Consumer { /* private fields */ }

Subscribe-style consumer handle. Construct via Consumer::builder.

Implementations

impl Consumer

pub async fn commit_sync(&self) -> Result<(), ConsumerError>

Commit the current next-offsets for every assigned partition. Blocks until the broker acks.

pub fn commit_async(&self)

Fire-and-forget commit. Returns once the request is enqueued on the client’s writer task; does NOT wait for the broker ack. Errors are logged but not returned.

impl Consumer

pub async fn start( bootstrap: String, client_id: String, group_id: String, session_timeout: Duration, rebalance_timeout: Duration, heartbeat_interval: Duration, subscribe: Vec<String>, auto_offset_reset: AutoOffsetReset, isolation_level: IsolationLevel, assignor: Assignor, client_rack: Option<String>, security: Option<ClientSecurity>, ) -> Result<Self, ConsumerError>

Build a Consumer subscribed to the given topics: resolve bootstrap, JoinGroup (twice), compute the assignment if we’re the elected leader, SyncGroup, prime offsets, then spawn the coordinator task that owns the heartbeat + rebalance loop.

pub fn builder() -> ConsumerStartBuilder

Build a Consumer subscribed to the given topics: resolve bootstrap, JoinGroup (twice), compute the assignment if we’re the elected leader, SyncGroup, prime offsets, then spawn the coordinator task that owns the heartbeat + rebalance loop.

impl Consumer

pub fn group_id(&self) -> &str

The consumer’s group id.

pub fn member_id(&self) -> &str

The member id assigned by the coordinator at join time.

pub fn generation_id(&self) -> i32

The generation id captured at the most recent successful join.

pub fn group_metadata(&self) -> ConsumerGroupMetadata

KIP-447 group metadata to hand to a transactional producer’s send_offsets_to_transaction. The generation id is the value captured at the most recent successful join (the field is not kept in sync as the coordinator rejoins — see Self::generation_id); for a stable single-member group this equals the coordinator’s live generation. group_instance_id is always None — the consumer has no static-membership support yet.

pub fn subscribed_topics(&self) -> &[String]

Topics this consumer subscribed to at build time.

pub async fn assignment(&self) -> Vec<(String, i32)>

Snapshot of currently assigned (topic, partition) pairs.

pub async fn close(self) -> Result<(), ConsumerError>

Stop the coordinator task so the broker evicts this member promptly.

The coordinator itself sends a best-effort LeaveGroup as the last thing it does on shutdown (see crate::coordinator::run), using its live member_id. That id can differ from the one captured at build time — a from-scratch rejoin (UNKNOWN_MEMBER_ID) replaces it — so the leave must come from the coordinator, which owns the current value; sending it here with self.member_id would silently leave a stale id and orphan the real member until its session expires. Cancel + join is prompt because the coordinator races its in-tick RPCs against the shutdown token.

impl Consumer

pub async fn poll( &mut self, timeout: Duration, ) -> Result<Vec<ConsumerRecord>, ConsumerError>

Returns the records from every v2 batch the broker returned per assigned partition, or an empty vec on timeout. Under read_committed isolation, control batches and records belonging to aborted transactions are filtered client-side using the response’s aborted_transactions list (the broker returns verbatim bytes). Rebalances are handled transparently by the internal coordinator task, which mutates the live assigned snapshot in place; poll() simply reads it on each call.