:rocket: The journey continues! Part 4 of my "𝐆𝐞...

# feed

[Jaehyeon Kim]

🚀 The journey continues! Part 4 of my "𝐆𝐞𝐭𝐭𝐢𝐧𝐠 𝐒𝐭𝐚𝐫𝐭𝐞𝐝 𝐰𝐢𝐭𝐡 𝐑𝐞𝐚𝐥-𝐓𝐢𝐦𝐞 𝐒𝐭𝐫𝐞𝐚𝐦𝐢𝐧𝐠 𝐢𝐧 𝐊𝐨𝐭𝐥𝐢𝐧" series is here:

"𝐅𝐥𝐢𝐧𝐤 𝐃𝐚𝐭𝐚𝐒𝐭𝐫𝐞𝐚𝐦 𝐀𝐏𝐈 - 𝐒𝐜𝐚𝐥𝐚𝐛𝐥𝐞 𝐄𝐯𝐞𝐧𝐭 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐢𝐧𝐠 𝐟𝐨𝐫 𝐒𝐮𝐩𝐩𝐥𝐢𝐞𝐫 𝐒𝐭𝐚𝐭𝐬"!

Having explored the lightweight power of Kafka Streams, we now level up to a full-fledged distributed processing engine: 𝗔𝗽𝗮𝗰𝗵𝗲 𝗙𝗹𝗶𝗻𝗸. This post dives into the foundational DataStream API, showcasing its power for stateful, event-driven applications.

In this deep dive, you'll learn how to:

• Implement sophisticated event-time processing with Flink's native 𝐖𝐚𝐭𝐞𝐫𝐦𝐚𝐫𝐤𝐬.
• Gracefully handle late-arriving data using Flink’s elegant 𝐒𝐢𝐝𝐞 𝐎𝐮𝐭𝐩𝐮𝐭𝐬 feature.
• Perform stateful aggregations with custom 𝐀𝐠𝐠𝐫𝐞𝐠𝐚𝐭𝐞𝐅𝐮𝐧𝐜𝐭𝐢𝐨𝐧 and 𝐖𝐢𝐧𝐝𝐨𝐰𝐅𝐮𝐧𝐜𝐭𝐢𝐨𝐧.
• Consume Avro records and sink aggregated results back to Kafka.
• Visualize the entire pipeline, from source to sink, using 𝐊𝐩𝐨𝐰 and 𝐅𝐚𝐜𝐭𝐨𝐫 𝐇𝐨𝐮𝐬𝐞 𝐋𝐨𝐜𝐚𝐥.
This is post 4 of 5, demonstrating the control and performance you get with Flink's core API. If you're ready to move beyond the basics of stream processing, this one's for you!

Read the full article here: https://jaehyeon.me/blog/2025-06-10-kotlin-getting-started-flink-datastream/

In the final post, we'll see how Flink's Table API offers a much more declarative way to achieve the same result. Your feedback is always appreciated!

🔗 𝐂𝐚𝐭𝐜𝐡 𝐮𝐩 𝐨𝐧 𝐭𝐡𝐞 𝐬𝐞𝐫𝐢𝐞𝐬:
1. Kafka Clients with JSON
2. Kafka Clients with Avro
3. Kafka Streams for Supplier Stats