- Notifications
You must be signed in to change notification settings - Fork308
A fully asynchronous, futures-based Kafka client library for Rust based on librdkafka
License
fede1024/rust-rdkafka
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
A fully asynchronous,futures-enabledApache Kafka clientlibrary for Rust based onlibrdkafka.
rust-rdkafka
provides a safe Rust interface to librdkafka. This versionis compatible with librdkafka v1.9.2+.
The main features provided at the moment are:
- Support for all Kafka versions since 0.8.x. For more information aboutbroker compatibility options, check thelibrdkafkadocumentation.
- Consume from single or multiple topics.
- Automatic consumer rebalancing.
- Customizable rebalance, with pre and post rebalance callbacks.
- Synchronous or asynchronous message production.
- Customizable offset commit.
- Create and delete topics and add and edit partitions.
- Alter broker and topic configurations.
- Access to cluster metadata (list of topic-partitions, replicas, activebrokers etc).
- Access to group metadata (list groups, list members of groups, hostnames,etc.).
- Access to producer and consumer metrics, errors and callbacks.
- Exactly-once semantics (EOS) via idempotent and transactional producersand read-committed consumers.
rust-rdkafka
is designed to be easy and safe to use thanks to theabstraction layer written in Rust, while at the same time being extremelyfast thanks to the librdkafka C library.
Here are some benchmark results using theBaseProducer
,sending data to a single Kafka 0.11 process running in localhost (defaultconfiguration, 3 partitions). Hardware: Dell laptop, with Intel Corei7-4712HQ @ 2.30GHz.
Scenario: produce 5 million messages, 10 bytes each, wait for all of them to be acked
- 1045413 messages/s, 9.970 MB/s (average over 5 runs)
Scenario: produce 100000 messages, 10 KB each, wait for all of them to be acked
- 24623 messages/s, 234.826 MB/s (average over 5 runs)
For more numbers, check out thekafka-benchmark project.
rust-rdkafka
provides low level and high level consumers and producers.
Low level:
BaseConsumer
: a simple wrapper around the librdkafka consumer. Itmust be periodicallypoll()
ed in order to execute callbacks, rebalancesand to receive messages.BaseProducer
: a simple wrapper around the librdkafka producer. As inthe consumer case, the user must callpoll()
periodically to executedelivery callbacks.ThreadedProducer
: aBaseProducer
with a separate thread dedicated topolling the producer.
High level:
StreamConsumer
: aStream
of messages that takes care ofpolling the consumer automatically.FutureProducer
: aFuture
that will be completed oncethe message is delivered to Kafka (or failed).
For more information about consumers and producers, refer to theirmodule-level documentation.
Warning: the library is under active development and the APIs are likelyto change.
Tokio is a platform for fast processing of asynchronous events in Rust.The interfaces exposed by theStreamConsumer
and theFutureProducer
allow rust-rdkafka users to easily integrate Kafka consumers and producerswithin the Tokio platform, and write asynchronous message processing code.Note that rust-rdkafka can be used without Tokio.
To see rust-rdkafka in action with Tokio, check out theasynchronous processing example in the examples folder.
At-least-once delivery semantics are common in many streaming applications:every message is guaranteed to be processed at least once; in case oftemporary failure, the message can be re-processed and/or re-delivered,but no message will be lost.
In order to implement at-least-once delivery the stream processingapplication has to carefully commit the offset only once the message hasbeen processed. Committing the offset too early, instead, might causemessage loss, since upon recovery the consumer will start from the nextmessage, skipping the one where the failure occurred.
To see how to implement at-least-once delivery withrdkafka
, check out theat-least-once delivery example in the examples folder. To know more aboutdelivery semantics, check themessage delivery semantics chapter in theKafka documentation.
Exactly-once semantics (EOS) can be achieved using transactional producers,which allow produced records and consumer offsets to be committed or abortedatomically. Consumers that set theirisolation.level
toread_committed
will only observe committed messages.
EOS is useful in read-process-write scenarios that require messages to beprocessed exactly once.
To learn more about using transactions in rust-rdkafka, see theTransactions section of the producer documentation.
Here are some of the projects using rust-rdkafka:
- kafka-view: a web interface for Kafka clusters.
- kafka-benchmark: a high performance benchmarking tool for Kafka.
- callysto: Stream processing framework in Rust.
- bytewax: Python stream processing framework using Timely Dataflow.
If you are using rust-rdkafka, please let us know!
Add this to yourCargo.toml
:
[dependencies]rdkafka = {version ="0.25",features = ["cmake-build"] }
This crate will compile librdkafka from sources and link it statically toyour executable. To compile librdkafka you'll need:
- the GNU toolchain
- GNU
make
pthreads
zlib
: optional, but included by default (feature:libz
)cmake
: optional,not included by default (feature:cmake-build
)libssl-dev
: optional,not included by default (feature:ssl
)libsasl2-dev
: optional,not included by default (feature:gssapi
)libzstd-dev
: optional,not included by default (feature:zstd-pkg-config
)
Note that using the CMake build system, via thecmake-build
feature, isencouraged if you can take the dependency on CMake.
By default a submodule with the librdkafka sources pinned to a specificcommit will be used to compile and statically link the library. Thedynamic-linking
feature can be used to instead dynamically link rdkafka tothe system's version of librdkafka. Example:
[dependencies]rdkafka = {version ="0.25",features = ["dynamic-linking"] }
If you'd like to compile librdkafka statically yourself, then usethat, you can usestatic-linking
while supplyingDEP_LIBRDKAFKA_STATIC_ROOT
with path to where librdkafka was built.
For a full listing of features, consult therdkafka-sys crate'sdocumentation. All of rdkafka-sys features arere-exported as rdkafka features.
The current minimum supported Rust version (MSRV) is 1.70.0. Note thatbumping the MSRV is not considered a breaking change. Any release ofrust-rdkafka may bump the MSRV.
Some features of theStreamConsumer
andFutureProducer
depend onTokio, which can be a heavyweight dependency for users who only intend touse the low-level consumers and producers. The Tokio integration isenabled by default, but can be disabled by turning off default features:
[dependencies]rdkafka = {version ="0.25",default-features =false }
If you would like to use an asynchronous runtime besides Tokio, you canintegrate it with rust-rdkafka by providing a shim that implements theAsyncRuntime
trait. See the following examples for details:
You can find examples in theexamples
folder. To run them:
cargo run --example<example_name> --<example_args>
rust-rdkafka uses thelog
crate to handle logging.Optionally, enable thetracing
feature to emittracing
events as opposed tolog
records.
In test and examples, rust-rdkafka uses theenv_logger
crateto format logs. In those contexts, logging can be enabledusing theRUST_LOG
environment variable, for example:
RUST_LOG="librdkafka=trace,rdkafka::client=debug" cargotest
This will configure the logging level of librdkafka to trace, and the levelof the client module of the Rust client to debug. To actually receive logsfrom librdkafka, you also have to set thedebug
option in the producer orconsumer configuration (see librdkafkaconfiguration).
To enable debugging in your project, make sure you initialize the loggerwithenv_logger::init()
, or the equivalent for anylog
-compatiblelogging framework.
Seerdkafka-sys.
Thanks to:
- Thijs Cadier -thijsc
- kafka-rust: a pure Rust implementation of the Kafka client.
About
A fully asynchronous, futures-based Kafka client library for Rust based on librdkafka
Topics
Resources
License
Uh oh!
There was an error while loading.Please reload this page.
Stars
Watchers
Forks
Packages0
Uh oh!
There was an error while loading.Please reload this page.