<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Julien Chanaud - Streemlined.io on Medium]]></title>
        <description><![CDATA[Stories by Julien Chanaud - Streemlined.io on Medium]]></description>
        <link>https://medium.com/@streemlined.io?source=rss-9756d4271f1c------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*H_g9_vQh70umgKcNSTkfrw.png</url>
            <title>Stories by Julien Chanaud - Streemlined.io on Medium</title>
            <link>https://medium.com/@streemlined.io?source=rss-9756d4271f1c------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 20:25:35 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@streemlined.io/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Building Kafka apps is way too complicated for the masses]]></title>
            <link>https://medium.com/@streemlined.io/building-kafka-apps-is-way-too-complicated-for-the-masses-be7a162800da?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/be7a162800da</guid>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[kafka-connect]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Tue, 16 Jun 2026 13:07:38 GMT</pubDate>
            <atom:updated>2026-06-16T13:07:38.834Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/867/1*QrySbr9x3LKSUlTkkpdROA.png" /></figure><p>I believe the current state of “building Kafka apps” is way too complicated for the masses.</p><p>I’ve experienced this first hand in my early Kafka days in 2019 when I needed to deploy and monitor multiple components just to be able to massage and move data from Kafka to a database.</p><p>I was skeptical about the “event streaming” ideas and I still am. I think Confluent has tried hard to sell this on every use case possible where all that was really necessary in 80% cases was “just a dumb integration”.</p><p>Not dissimilar to every vendor trying to shove AI down your throat today.</p><p>Integration on the other hand, is something that has become harder with Kafka.</p><p>You can criticize the old Talend, Datastage, Informatica and the likes all you want, they come from a different era, no doubt. Their deployment and operating models are not relevant anymore. No more centralized brokers and application servers. Good.</p><p>But they were good at something:</p><p>Developer Experience.</p><p>I spent the past 6 months building Streemlined because I believe Kafka needs the Developer Experience of Talend with the operating model of Kafka Streams.</p><p>Come talk.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=be7a162800da" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You have simple Kafka use cases, yet use the heavy artillery every single time]]></title>
            <link>https://medium.com/@streemlined.io/you-have-simple-kafka-use-cases-yet-use-the-heavy-artillery-every-single-time-4d71f6c1cfb8?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/4d71f6c1cfb8</guid>
            <category><![CDATA[apache-kafka]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Mon, 08 Jun 2026 11:02:28 GMT</pubDate>
            <atom:updated>2026-06-08T11:02:28.240Z</atom:updated>
            <content:encoded><![CDATA[<p>You have simple Kafka use cases, and yet you’ve been told to use the heavy artillery every single time.</p><p>Sink a JSON topic in database <br>➡️Use Kafka Streams + Kafka Connect (can’t use schemaless with JdbcSink)</p><p>Sink an Avro topic in database, but filter out unrelevant records? <br>➡️Kafka Streams + Kafka Connect (or use SMTs)</p><p>Anything slightly trickier? <br>➡️Use Flink</p><p>2026?<br>➡️Use Agentic</p><p>None of these suggestions make much sense for an application team.</p><p>A bazooka to kill a fly.</p><p>Kafka isn’t hard. <br>The ecosystem built around it is just made for a difference audience.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/276/1*HbX915iqrBGjOvCMhaINbA.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4d71f6c1cfb8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[You’re a solo founder, that’s a risk]]></title>
            <link>https://medium.com/@streemlined.io/youre-a-solo-founder-that-s-a-risk-5419795bd622?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/5419795bd622</guid>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Thu, 04 Jun 2026 14:08:58 GMT</pubDate>
            <atom:updated>2026-06-04T14:08:58.725Z</atom:updated>
            <content:encoded><![CDATA[<p>Last week I vented about the paradox:</p><blockquote>“You’re a solo founder, that’s a risk”</blockquote><p>… said by people happily building on a library maintained by Dave during his free time.</p><p>This week I come with answers instead:</p><p>𝗣𝗲𝗿𝗽𝗲𝘁𝘂𝗮𝗹 𝗿𝘂𝗻 𝗹𝗶𝗰𝗲𝗻𝘀𝗲<br>Your Kafka app keeps running forever. No kill switch.</p><p>𝗦𝗼𝘂𝗿𝗰𝗲 𝗰𝗼𝗱𝗲 𝗲𝘀𝗰𝗿𝗼𝘄<br>If I go dark for 90 days, you get the source.</p><p>𝗢𝗽𝗲𝗻-𝘀𝗼𝘂𝗿𝗰𝗲<br>If I ever stop, the whole codebase goes public. Fork it and keep going.</p><p><a href="https://www.linkedin.com/company/streemlined/">Streemlined</a> has everything that makes Dave’s repo feel safe: the code, the freedom, the exit.<br>Plus a founder who treats you like a tier-1 priority.</p><p>Continuity isn’t a clause in a contract. It’s how I write software.</p><p>👉 <a href="https://www.streemlined.io/our-promise">https://www.streemlined.io/our-promise</a></p><p>(no offense Dave, I love your work)</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5419795bd622" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The ultimate corporate paradox]]></title>
            <link>https://medium.com/@streemlined.io/the-ultimate-corporate-paradox-88ec30097020?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/88ec30097020</guid>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Thu, 04 Jun 2026 14:07:11 GMT</pubDate>
            <atom:updated>2026-06-04T14:07:11.524Z</atom:updated>
            <content:encoded><![CDATA[<blockquote>“We can’t buy your software because you’re a solo founder, and that’s a risk.”</blockquote><p>So instead, they go and build on top of an open-source GitHub library maintained by… one guy named Dave in his free time between 11 PM and 2 AM.</p><p>Somehow, the hobbyist is a “safe bet”, but the dedicated founder treating them like a tier-1 priority is a “liability”.</p><p>Make it make sense. 🤔</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=88ec30097020" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Python & JS transform for Kafka Connect]]></title>
            <link>https://medium.com/@streemlined.io/python-js-transform-for-kafka-connect-445570b84d9e?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/445570b84d9e</guid>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[kafka-connect]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Wed, 03 Jun 2026 13:28:11 GMT</pubDate>
            <atom:updated>2026-06-03T13:28:11.251Z</atom:updated>
            <content:encoded><![CDATA[<p>Following last week’s discussions at <a href="https://www.linkedin.com/search/results/all/?keywords=%23current26&amp;origin=HASH_TAG_FROM_FEED">Current 26</a> London, I’m currently experimenting how to make use of GraalVM Polyglot API in <a href="https://www.linkedin.com/company/streemlined/">Streemlined</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wYsqCrtkleNJTl_68u9HMA.png" /></figure><p>What you see here is a new node, “Code Transform”, which lets you convert any input message into any output message using Python or JavaScript.</p><p>- Test your code directly against sample messages.<br>- Automatically verify that the result conforms to the output schema.<br>- As usual, you only worry about the business logic. Error handling, observability, … are all taken care of by Streemlined.</p><p>Performance wise, it takes less than 10 microseconds to process this simple code. We’re in the 100k msg/s range, we’re good.</p><p>I really like the JSONata transform, but I must admit Python and JS are more intuitive for the majority of people.</p><p>At this point, nothing’s stopping me from implementing state stores.</p><p>Fancy a RocksDB persistent state store backed by a Kafka topic? <br>Come talk, let’s do it.</p><p>Can’t wait to demo a word count app. <br>Or a fraud detection app. <br>lol</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=445570b84d9e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[IBM MQ and MongoDB in Streemlined!]]></title>
            <link>https://medium.com/@streemlined.io/ibm-mq-and-mongodb-in-streemlined-57a15196b458?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/57a15196b458</guid>
            <category><![CDATA[ibm]]></category>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[mongodb]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Wed, 03 Jun 2026 13:25:42 GMT</pubDate>
            <atom:updated>2026-06-03T13:25:42.066Z</atom:updated>
            <content:encoded><![CDATA[<p>A few days ago I demonstrated the Streemlined AI pipeline builder.<br>If you haven’t seen it, go check it out: I’m pouring my coffee while the thing builds itself. Pretty cool!</p><p>All the features and user experience I’ve been implementing in <a href="https://www.linkedin.com/company/streemlined/">Streemlined</a> to make building Kafka application more accessible for the non-experts also happen to be the exact same tools an LLMs needs.<br>A user will drag the required nodes on the canvas, configure them and fix the validation errors as they appear.<br>That feedback loop works the same way whether you’re a human or an LLM.</p><p>A few weeks ago I demonstrated how you can use ANY Source or Sink Connector from the Kafka ecosystem inside Streemlined.</p><p>Something I didn’t mention is that not all Source Connectors produce a record with a schema attached. That is problematic because the entire user experience and validation inside Streemlined relies on the presence of schemas. If you’re missing a schema at the source, you have to define it manually and that can be a source of errors.<br>Both for humans and LLMs.</p><p>And that’s why I’m extremely proud to introduce a new schema inference module for all the Kafka Connect Source Connectors.<br>Once you’ve configured your Source, just click one button and Streemlined will “run” the task once and automatically infer the schema from the messages it polled. It works irrespective of the Source Connector record. If it’s a Struct, great, we already have the Schema, but if it’s a JSON-ish String, it’ll work the same!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yklmqdrQXs5-MfvQZZN8Dg.png" /></figure><p>In these examples I’m configuring two different connectors, IBM MQ Source and MongoDB Source. In a matter of seconds, I get a schema proposition, which I can amend and accept.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L8QLoSOFE75gpqHeHK4djQ.png" /></figure><p>What’s even more interesting is that from now on, poison pills (that is, messages that don’t conform to that schema) can be dealt with however you see fit. Maybe you want to sent them to a DLQ, maybe you want to fail the pipeline.</p><p>Streemlined gives you options, inside the canvas, so you don’t need to reinvent the wheel.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=57a15196b458" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Voice to fully working Kafka application in less than 3 minutes.]]></title>
            <link>https://medium.com/@streemlined.io/voice-to-fully-working-kafka-application-in-less-than-3-minutes-bb20933758b0?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/bb20933758b0</guid>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[kafka-connect]]></category>
            <category><![CDATA[kafka-streams]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Tue, 12 May 2026 09:08:18 GMT</pubDate>
            <atom:updated>2026-05-12T09:08:18.000Z</atom:updated>
            <content:encoded><![CDATA[<p>Do you believe building Kafka applications could be made accessible to non-developpers? I hope this new demo will convince you.</p><p>I’ve spent the past few months building <a href="https://www.linkedin.com/company/streemlined/">Streemlined</a>, focused on lowering the barrier to building Kafka application by providing the best possible user experience.</p><p>I was recently challenged that despite all my efforts, the product still requires some form of expertise. “Who wants to read a stack trace?”<br>So I reworked the validation engine entirely.</p><p>Now, the product informs you in real time about what needs fixing while you’re building your pipeline. <br>An invalid value.<br>A missing required configuration.<br>An incompatible schema between two nodes.</p><p>You draw the general layout of your application, and you fix the messages that appear along the way.</p><p>That feedback loop brings the User Experience to a completely new level and makes building applications with Streemlined truly enjoyable.</p><p>Could we now apply that same Act &amp; Observe feedback loop to an LLM?</p><p>Enjoy the 2 minutes video.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FzL5DbsJBXO8%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fshorts%2FzL5DbsJBXO8&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FzL5DbsJBXO8%2Fhq2.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/5e7e0f99247fc476df8aa86d084a4890/href">https://medium.com/media/5e7e0f99247fc476df8aa86d084a4890/href</a></iframe><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bb20933758b0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Kafka Connect, but better!]]></title>
            <link>https://medium.com/@streemlined.io/kafka-connect-but-better-657c294b2429?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/657c294b2429</guid>
            <category><![CDATA[kafka-streams]]></category>
            <category><![CDATA[data-integration]]></category>
            <category><![CDATA[kafka-connect]]></category>
            <category><![CDATA[apache-kafka]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Wed, 29 Apr 2026 21:04:05 GMT</pubDate>
            <atom:updated>2026-04-29T21:04:05.710Z</atom:updated>
            <content:encoded><![CDATA[<p>You wanted to build a simple integration between another team’s Kafka topic and your database. You ended up with weeks of infrastructure headaches and platform team bottlenecks.</p><p>On paper, it looks easy: use Kafka Connect right?<br>Well… Not quite!</p><p>Your organization have applied the best practices, you have meaningful business events with rich details and structure to them. <br>They don’t fit straight into a table.</p><p>Now here are your options:</p><p>If you go All-in on Kafka Connect, the moment a basic SMT isn’t enough, you’ll be forced to write custom JARs and beg the Platform Team to restart a shared cluster just to accomodate a single integration’s requirements.<br>And you will end up with the most ugly and untestable piece of JSON configuration on the planet.</p><p>If you choose Kafka Streams, you’re told it’s “best practice” to never touch a database. So now you need both: A Kafka Streams app for the logic and a Connect cluster for the egress. Two different tech stacks with completely different operating models, a mountain of boilerplate, and a distributed system nightmare just to save a row to a table.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*eczEzi8r8JiRr58Y" /></figure><p><a href="https://www.linkedin.com/company/streemlined/">Streemlined</a> 𝘄𝗮𝘀 𝗯𝘂𝗶𝗹𝘁 𝗯𝗲𝗰𝗮𝘂𝘀𝗲 𝘆𝗼𝘂 𝘀𝗵𝗼𝘂𝗹𝗱𝗻’𝘁 𝗵𝗮𝘃𝗲 𝘁𝗼 𝗰𝗵𝗼𝗼𝘀𝗲 𝗯𝗲𝘁𝘄𝗲𝗲𝗻 𝗮 𝗿𝗶𝗴𝗶𝗱 𝗿𝘂𝗻𝘁𝗶𝗺𝗲 𝗮𝗻𝗱 𝗮 𝗽𝗶𝗰𝗸𝘆 𝗽𝗿𝗶𝗻𝗰𝗲𝘀𝘀.<br>It combines the connectivity of Connect with the transformation power of Streams into a single, visual workflow.<br>- No More SMT JARs: Perform complex flattening, filtering, and masking visually.<br>- Integration-First: Unlike pure Streams, Streemlined is built to talk to your infrastructure directly. JDBC, MQ, MongoDB… anything!<br>- Total Autonomy: Every pipeline compiles into its own standalone container. No shared clusters, no platform team bottlenecks.</p><p>If you’re looking to trial a container-first approach that respects your time, get in touch. <br><a href="https://www.linkedin.com/company/streemlined/">Streemlined</a> can help you.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=657c294b2429" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[POV: Optimizing GitHub Actions]]></title>
            <link>https://medium.com/@streemlined.io/pov-optimizing-github-actions-be175c4a0cae?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/be175c4a0cae</guid>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Wed, 29 Apr 2026 21:02:00 GMT</pubDate>
            <atom:updated>2026-04-29T21:02:00.164Z</atom:updated>
            <content:encoded><![CDATA[<p>Being a Product Manager for several years, I became expert at saying “No” to absolutely everybody.</p><p>I said No to my engineers.<br>I said No to my customers.<br>I said No to Sales. To Customer Success.<br>I especially 😂 said No to my boss.</p><p>If you wanted to get a Yes, you better come prepared. You had to prove that your request actually moved the needle for the company before it earned a spot on the roadmap.</p><p>And yet, here I am spending 4 hours optimizing GitHub Actions to shave off 1 minute out of a 3 minutes build.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/802/0*CIBmreXb_B9kpk2k" /></figure><p>Nodoby can stop me from prematurely optimizing a process that will not recoup in a lifetime.</p><p>Instead of compiling a list of use cases for <a href="https://www.linkedin.com/feed/#"><strong>Streemlined</strong></a> for example. Something I’ve been told to do several times, by multiple people.</p><p>I’ve finally met the one stakeholder I can’t easily shoot down.</p><p>Myself.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=be175c4a0cae" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Troubleshooting Confluent MQ Source Connector]]></title>
            <link>https://medium.com/@streemlined.io/troubleshooting-confluent-mq-source-connector-384823798e53?source=rss-9756d4271f1c------2</link>
            <guid isPermaLink="false">https://medium.com/p/384823798e53</guid>
            <category><![CDATA[kafka-connect]]></category>
            <category><![CDATA[apache-kafka]]></category>
            <category><![CDATA[kafka]]></category>
            <category><![CDATA[ibm]]></category>
            <dc:creator><![CDATA[Julien Chanaud - Streemlined.io]]></dc:creator>
            <pubDate>Fri, 17 Apr 2026 21:55:04 GMT</pubDate>
            <atom:updated>2026-04-17T21:55:04.665Z</atom:updated>
            <content:encoded><![CDATA[<p>A prospect reported that the Confluent MQ Source Connector didn’t work with <a href="https://www.streemlined.io/">Streemlined</a>.</p><p>I love a good challenge so I quickly downloaded the connector into Streemlined plugins folder, launched an IBM MQ server and got to debugging.</p><p>After configuring the connector properly in Streemlined:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DYCrDGktc14E35_0VYrKXw.png" /></figure><p>I got the following exception while trying to start the application:</p><pre>java.lang.NoSuchMethodError: &#39;void org.apache.kafka.connect.util.KafkaBasedLog.&lt;init&gt;(java.lang.String, java.util.Map, java.util.Map, org.apache.kafka.connect.util.Callback, org.apache.kafka.common.utils.Time, java.lang.Runnable)&#39;<br>	at io.confluent.license.LicenseStore.createKafkaBasedLog(LicenseStore.java:225)<br>	...</pre><p>Interesting, KafkaBasedLog is part of the Apache Kafka Connect runtime. Confluent relies on this to perform its License Check.</p><p>Looking into the <a href="https://github.com/apache/kafka/commit/982ea2f6a471c217c7400a725c9504d9d8348d02#diff-b807bff8df435f958053a508456fe1df8769110c45d5359ba9ad5af04388764eR246">Github commit history</a>, we can see this constructor for KafkaBasedLog was flagged as <strong>deprecated </strong>in Apache Kafka 3.0. <br>In 2021!</p><p>And last year, the deprecated constructor <a href="https://github.com/apache/kafka/commit/e9d4aa4cfeeecb56eac9b9543d5d403b691f1768">finally got removed</a> with Kafka 4.0</p><p>Confluent License check still relies on this 5 years old deprecated code.</p><blockquote>But Julien, it works in my company without any issue!</blockquote><p>Yes! It does, because you’ve deployed the Confluent proprietary jars, namely connect-runtime-*-ce.jar or connect-runtime-*-ccs.jar. Those jars must still include the deprecated constructor.</p><p>To Confluent’s defense, it’s not such a big deal because if you’re paying for Confluent licensed connectors, it’s just reasonable that you also happen to deploy them on the Confluent build of Kafka Connect.</p><p>Anyway, let’s bring the missing constructor back! By performing some Java trickery, namely class shadowing, we should be fine…</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*cumbrPijV7LvUY88vqEu6g.png" /><figcaption>Class shadowing KafkaBasedLog class: you just drop the class in your src folder so it gets picked first by the classloader</figcaption></figure><p>Let’s try again! And… still no success, but progress!</p><pre>java.lang.NoSuchMethodError: &#39;org.apache.kafka.connect.util.TopicAdmin$NewTopicBuilder org.apache.kafka.connect.util.TopicAdmin$NewTopicBuilder.minInSyncReplicas(short)&#39;<br>	at io.confluent.license.LicenseStore.createTopicDescription(LicenseStore.java:428)<br>        ...</pre><p>This time, Confluent’s license check is trying to call a function with the signature minInSyncReplicas(short) in <a href="https://github.com/apache/kafka/blob/trunk/connect/runtime/src/main/java/org/apache/kafka/connect/util/TopicAdmin.java">org.apache.kafka.connect.util.TopicAdmin$NewTopicBuilder</a> class.</p><p>Confluent must have updated the Apache Kafka code directly, but didn’t send the patch upstream. Ok, I can do that too!</p><p>Those little things make you wonder how different Confluent Kafka is compared to Apache Kafka. <br>Slightly different, for sure. But we’re not allowed to decompile the Confluent jars to figure out “by how much” so we’ll have to stick to just reading the logs and guessing.</p><p>In any case, let’s implement the missing function.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9TGnPTTZG3CWMwc4gtKbsA.png" /></figure><p>Let’s try again:</p><pre>INFO  i.c.c.i.m.IbmMQSourceConnectorConfig - Attempting to create IBM MQ connection factory with configuration.<br>INFO  i.c.c.i.m.IbmMQSourceConnectorConfig - Created IBM MQ connection factory with configuration.<br>INFO  i.c.c.j.core.source.MessageProcessor - Trying to (re)establish connection with connection 1<br>INFO  i.c.c.jms.core.source.JmsConnection - Setting JMS client ID for connection connection 1 to &#39;null-0394d751c&#39;<br>INFO  i.c.c.jms.core.source.JmsConnection - Creating JMS consumer 1 for session connection 1<br>INFO  i.c.c.jms.core.source.JmsConnection - Created consumer.<br>INFO  i.c.c.j.core.source.MessageProcessor - Successfully (re)established connection with connection 1</pre><p>Success!</p><p>And after a playing around a little bit with Streemlined, I managed to implement a small IBM MQ to Postgres prototype, just like that.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HPUGUAW7KN0YuKmOMWp5-A.png" /></figure><p>Quite funny to be pushing the Kafka Connect model so far that it doesn’t even need Kafka anymore.</p><p>If you’re interested to see how Streemlined can help you speed up building Kafka applications while making them more accessible, get in touch!</p><p>Hats off to the IBM MQ team in charge of MQ for docker. I only had to delpoy this tiny compose:</p><pre>services: <br>  mq:<br>    image: icr.io/ibm-messaging/mq<br>    hostname: ibmmq<br>    environment:<br>      LICENSE: &quot;accept&quot;<br>      MQ_QMGR_NAME: &quot;QM1&quot;<br>      MQ_ADMIN_PASSWORD: &quot;123456789&quot;<br>      MQ_APP_PASSWORD: &quot;123456789&quot;<br>    ports:<br>      - &quot;1414:1414&quot;<br>      - &quot;9443:9443&quot;</pre><p>to get the benefit of:</p><ul><li>a fully functionning IBM MQ dev server</li><li>multiple Queues already created</li><li>2 users (admin and app)</li><li>a rudimentary web interface to get and put messages</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lO1GZqRmqds8Y2nI5Osxng.png" /></figure><p>I was geniunely scared I was going to need <strong>rfhutil</strong>, can’t express how happy I was that the docker image was carefully considered for developer experience.</p><p>Thank you.</p><p>Other middlewares should take example from you there.</p><p>::wink-wink::</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=384823798e53" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>