🚨New Video Alert!🚨
Giving back to #openSource can be intimidating. You may think your code needs to be perfect or your idea be fully fleshed out. But the ideal contributor might surprise you...
@gharris1727 offers insights from the perspective of the #apacheKafka project. 👇
If you hate data loss bugs, check out my #current24 talk #ApacheKafka "MirrorMaker 2’s Offset Translation Isn’t Exactly Once (and That’s Okay)" https://t.co/Pk4vRutzMB
If you're interested in Apache #Kafka Cross-Cluster Replication, I want to hear from you at #KafkaSummit! Read KIP-986: https://t.co/m8rpZIFXSv and come find me at the event.
KIP-898: Modernize #KafkaConnect plugin discovery is released in #ApacheKafka 3.6.0: https://t.co/mLvZ5g1HHP
May your migrations be easy and your startups fast!
Excited to show you a #KafkaConnect IDE Plugin Prototype at #Current23 tomorrow! "A New UI for Kafka Connect: Your Favorite IDE!" https://t.co/Y65TQWi3HH
🎉 Excited to announce that my talk, "Exactly-Once, Again: Adding EOS support for Kafka Connect Source Connectors" has been accepted for Kafka Summit London 2023! See you all in May 🍻
@ijuma But just for you, some early numbers: Scanning went from 1s to 40ms on an empty Connect worker. And for reference, scanning on large plugin paths is currently in the 10s of seconds.
Faster startup and lower memory usage for Kafka Connect with KIP-898: Modernize Connect plugin discovery! Faster test iteration, more efficient production environments, and enables new use-cases. Read the KIP here: https://t.co/MPJ6AIDDyr
@ijuma I will include benchmarks once I have built the migration script, and can test against representative public connectors. The claims are based on the fact that strictly less work is being done by the ServiceLoader as compared to Reflections, and experience with the status quo.
@gunnarmorling I think that is where Connect must go in the future, in some form. It needs both the 'just-add-nodes' distributed model, and a container-ready mode for those that want to use external bin-packing and lifecycle management solutions, and not depend on our re-implementations.
A prototype for faster Kafka Connector configuration development: What if saving your configuration automatically deployed it locally?
More details and vision: https://t.co/CRO7Zu2Xmn
#kafka#kafkaconnect