Kafka logo

Kafka

Distributed, fault tolerant, high throughput pub-sub messaging system
22.9K
21.5K
+ 1
607

What is Kafka?

Kafka is a distributed, partitioned, replicated commit log service. It provides the functionality of a messaging system, but with a unique design.
Kafka is a tool in the Message Queue category of a tech stack.
Kafka is an open source tool with 27K GitHub stars and 13.4K GitHub forks. Here’s a link to Kafka's open source repository on GitHub

Who uses Kafka?

Companies
1513 companies reportedly use Kafka in their tech stacks, including Uber, Shopify, and Spotify.

Developers
20739 developers on StackShare have stated that they use Kafka.

Kafka Integrations

Datadog, Apache Flink, Presto, Databricks, and Couchbase are some of the popular tools that integrate with Kafka. Here's a list of all 104 tools that integrate with Kafka.
Pros of Kafka
126
High-throughput
119
Distributed
92
Scalable
86
High-Performance
66
Durable
38
Publish-Subscribe
19
Simple-to-use
18
Open source
12
Written in Scala and java. Runs on JVM
9
Message broker + Streaming system
4
KSQL
4
Avro schema integration
4
Robust
3
Suport Multiple clients
2
Extremely good parallelism constructs
2
Partioned, replayable log
1
Simple publisher / multi-subscriber model
1
Fun
1
Flexible
Decisions about Kafka

Here are some stack decisions, common use cases and reviews by companies and developers who chose Kafka in their tech stack.

Surabhi Bhawsar
Technical Architect at Pepcus · | 7 upvotes · 712.4K views
Needs advice
on
Apache FlinkApache Flink
and
KafkaKafka

I need to build the Alert & Notification framework with the use of a scheduled program. We will analyze the events from the database table and filter events that are falling under a day timespan and send these event messages over email. Currently, we are using Kafka Pub/Sub for messaging. The customer wants us to move on Apache Flink, I am trying to understand how Apache Flink could be fit better for us.

See more
Ashish Singh
Tech Lead, Big Data Platform at Pinterest · | 38 upvotes · 2.8M views

To provide employees with the critical need of interactive querying, we’ve worked with Presto, an open-source distributed SQL query engine, over the years. Operating Presto at Pinterest’s scale has involved resolving quite a few challenges like, supporting deeply nested and huge thrift schemas, slow/ bad worker detection and remediation, auto-scaling cluster, graceful cluster shutdown and impersonation support for ldap authenticator.

Our infrastructure is built on top of Amazon EC2 and we leverage Amazon S3 for storing our data. This separates compute and storage layers, and allows multiple compute clusters to share the S3 data.

We have hundreds of petabytes of data and tens of thousands of Apache Hive tables. Our Presto clusters are comprised of a fleet of 450 r4.8xl EC2 instances. Presto clusters together have over 100 TBs of memory and 14K vcpu cores. Within Pinterest, we have close to more than 1,000 monthly active users (out of total 1,600+ Pinterest employees) using Presto, who run about 400K queries on these clusters per month.

Each query submitted to Presto cluster is logged to a Kafka topic via Singer. Singer is a logging agent built at Pinterest and we talked about it in a previous post. Each query is logged when it is submitted and when it finishes. When a Presto cluster crashes, we will have query submitted events without corresponding query finished events. These events enable us to capture the effect of cluster crashes over time.

Each Presto cluster at Pinterest has workers on a mix of dedicated AWS EC2 instances and Kubernetes pods. Kubernetes platform provides us with the capability to add and remove workers from a Presto cluster very quickly. The best-case latency on bringing up a new worker on Kubernetes is less than a minute. However, when the Kubernetes cluster itself is out of resources and needs to scale up, it can take up to ten minutes. Some other advantages of deploying on Kubernetes platform is that our Presto deployment becomes agnostic of cloud vendor, instance types, OS, etc.

#BigData #AWS #DataScience #DataEngineering

See more
Joshua Dean Küpper
CEO at Scrayos UG (haftungsbeschränkt) · | 1 upvote · 175.8K views

We make extensive use of Redis for our caches and use it as a way to save "semi-permanent" stuff like user-submit settings (that get refreshed on each login) or cooldowns that expire very fast. Additionally we also utilize the Pub-Sub capabilities that Redis has to offer.

We decided against using a dedicated Message-Broker/Streaming Platform like RabbitMQ or Kafka, as we already had a packet-based, custom protocol for communication between servers and services, and we only needed some "tiny" Pub-Sub magic to fill in the gaps. An entire additional service just for this oddjob would've been a total overkill.

See more

I want to read data from Kafka. The file is in CSV, or whichever format is coming from SAP and read from another 3rd party application. So I need to create a Message Bus for the same. Please suggest. I can use a microservice as well.

See more

I have to build a data processing application with an Apache Beam stack and Apache Flink runner on an Amazon EMR cluster. I saw some instability with the process and EMR clusters that keep going down. Here, the Apache Beam application gets inputs from Kafka and sends the accumulative data streams to another Kafka topic. Any advice on how to make the process more stable?

See more
Needs advice
on
DjangoDjangoKafkaKafka
and
RedisRedis

I'm building a website where users can participate, like and dislike any given challenge.

Problem : If 10k or 1 million users join the given challenge at a time it can cause a race condition in my database MySQL and in also Redis.

What I want : Aggregating joined participated users, likes and dislikes.

Solution : I'm thinking about using Kafka as a Queue message broker then users event one by one saving into Redis, database and aggregate them.

One problem is also here saving and doing aggregate takes time now; how can I show users they have successfully joined the challenge?

One solution is that when a user joins the challenge I send a request to the Kafka queue then update the current user UI and show a success message (not updating the other users' joined messages to current user because I am not using Websockets)

Other App example Take the same example of https://stackshare.io posts. On posts users can like, dislike and comments.

Estimated users : 1 million Stack : Django, Mysql, Redis and Kafka

Questions

  • How I can manage these kinds of things?
  • How do big tech companies handle this?
  • Where am I right or wrong?
  • Are there other tools that can help me in this situation?
  • I am using locks in Redis when total like, dislike and joined users increment or decrement. Should I be doing this? Is it the same for transactions in MySQL?

I need the best approach to handle this situation that can also be scalable.

Thanks in advance for reading my post and giving me suggestions on this. ☺️

See more

Blog Posts

Dec 22 2021 at 5:41AM

Pinterest

MySQLKafkaDruid+3
3
560
Amazon S3KafkaZookeeper+5
8
1554
Mar 24 2021 at 12:57PM

Pinterest

GitJenkinsKafka+7
3
2119

Kafka's Features

  • Written at LinkedIn in Scala
  • Used by LinkedIn to offload processing of all page and other views
  • Defaults to using persistence, uses OS disk cache for hot data (has higher throughput then any of the above having persistence enabled)
  • Supports both on-line as off-line processing

Kafka Alternatives & Comparisons

What are some alternatives to Kafka?
ActiveMQ
Apache ActiveMQ is fast, supports many Cross Language Clients and Protocols, comes with easy to use Enterprise Integration Patterns and many advanced features while fully supporting JMS 1.1 and J2EE 1.4. Apache ActiveMQ is released under the Apache 2.0 License.
RabbitMQ
RabbitMQ gives your applications a common platform to send and receive messages, and your messages a safe place to live until received.
Amazon Kinesis
Amazon Kinesis can collect and process hundreds of gigabytes of data per second from hundreds of thousands of sources, allowing you to easily write applications that process information in real-time, from sources such as web site click-streams, marketing and financial information, manufacturing instrumentation and social media, and operational logs and metering data.
Apache Spark
Spark is a fast and general processing engine compatible with Hadoop data. It can run in Hadoop clusters through YARN or Spark's standalone mode, and it can process data in HDFS, HBase, Cassandra, Hive, and any Hadoop InputFormat. It is designed to perform both batch processing (similar to MapReduce) and new workloads like streaming, interactive queries, and machine learning.
Akka
Akka is a toolkit and runtime for building highly concurrent, distributed, and resilient message-driven applications on the JVM.
See all alternatives

Kafka's Followers
21505 developers follow Kafka to keep up with related blogs and decisions.