MERGE and Liquid Clustering: Common Performance Issues
(Wed, 21 Jan 2026)
As a Spark support engineer, I still encounter many cases where MERGE or JOIN operations on Delta tables do not perform as expected, even when liquid clustering is used. While liquid clustering
is a significant improvement over traditional partitioning and offers many advantages, people still sometimes struggle with it. There is often an assumption that enabling liquid clustering will
automatically result in efficient merges, but in practice, this is not always true, and the reason is a lack of understanding.
Here are the most common issues when executing a merge on a liquid clustering table.
>> Read More
Why High-Availability Java Systems Fail Quietly Before They Fail Loudly
(Wed, 21 Jan 2026)
Most engineers imagine failures as sudden events. A service crashes. A node goes down. An alert fires, and everyone jumps into action. In real high-availability Java systems, failures rarely
behave that way. They almost always arrive quietly first.
Systems that have been running reliably for months or years begin to show small changes. Latency creeps up. Garbage collection pauses last a little longer. Thread pools spend more time near
saturation. Nothing looks broken, and dashboards stay mostly green. Then one day, the system tips over, and the failure suddenly looks dramatic.
>> Read More
Multimodal AI Architecture: Unifying Vision, Text, and Sensor Intelligence
(Wed, 21 Jan 2026)
Most Android AI features today are still single-modal
A camera screen that does object detection.
A chat screen that calls an LLM.
A sensor-driven feature that runs in the background.
The real fun starts when you combine these: camera, text, sensors, history, and context. That’s where multimodal AI shines — and where architecture makes or breaks your app.
>> Read More
AI-Driven Autonomous CI/CD for Risk-Aware DevOps
(Wed, 21 Jan 2026)
Currently, the software development process relies on integrating development and operations (DevOps) to accelerate delivery without compromising quality. When the system becomes very complex, it
becomes risky and delays the manual control of the continuous integration or continuous deployment (CI/CD) processes. AI-based autonomous pipelines manage the entire process by automating
decisions, optimizing, and eliminating human errors.
Continuous risk-aware DevOps involves monitoring and signaling issues, as well as predicting failures. The self-healing mechanisms handle the whole thing in a way that minimises disruption and
improves system stability across different deployments.
>> Read More
DevOps Cafe Ep 79 - Guests: Joseph Jacks and Ben Kehoe
(Mon, 13 Aug 2018)
Triggered by Google Next 2018, John and Damon chat with Joseph Jacks (stealth startup) and Ben Kehoe (iRobot) about their public disagreements — and agreements — about Kubernetes and
Serverless.
>> Read More
DevOps Cafe Ep 78 - Guest: J. Paul Reed
(Mon, 23 Jul 2018)
John and Damon chat with J.Paul Reed (Release Engineering Approaches) about the field of Systems Safety and Human Factors that studies why accidents happen and how to minimize the occurrence and
impact.
Show notes at http://devopscafe.org
>> Read More
DevOps Cafe Ep. 77 - Damon interviews John
(Wed, 20 Jun 2018)
A new season of DevOps Cafe is here. The topic of this episode is "DevSecOps." Damon interviews John about what this term means, why it matters now, and the overall state of security.
Show notes at http://devopscafe.org
>> Read More