<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Streaming on nanta - Data Engineering</title><link>https://nanta-data.dev/en/tags/streaming/</link><description>Recent content in Streaming on nanta - Data Engineering</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>© 2026 nanta</copyright><lastBuildDate>Tue, 03 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://nanta-data.dev/en/tags/streaming/index.xml" rel="self" type="application/rss+xml"/><item><title>Flink on EKS In-place Scaling: Scaling TaskManagers Without Restarting the Job</title><link>https://nanta-data.dev/en/posts/flink-in-place-scaling/</link><pubDate>Tue, 03 Mar 2026 00:00:00 +0000</pubDate><guid>https://nanta-data.dev/en/posts/flink-in-place-scaling/</guid><description>Our recommendation system&amp;rsquo;s Flink application required sub-1-minute latency, which prevented us from using autoscaling or spot instances. Autoscaling or spot reclamation triggered full Flink restarts that took 2-3 minutes. Using Flink 1.18&amp;rsquo;s adaptive scheduler and K8s Operator 1.8, we enabled in-place scaling — reducing consumer lag peaks to 1/5 and lag duration from 5-7 minutes to 2-3 minutes during scale events.</description></item></channel></rss>