kloia Blog

Knowledge Base vs Knowledge Graph for LLM Systems (2026 Guide) | Kloia

Yagmur Akarken — Thu, 16 Apr 2026 21:07:03 GMT

Large language models contain a surprising amount of factual information baked into their parameters during pre-training. Ask GPT-4 who discovered Radium and it will answer correctly. Ask it what the capital of France is and it will not hesitate. This parametric knowledge is impressive, but it comes with three fundamental limitations that make it insufficient for production systems.

First, knowledge becomes outdated. A model trained on data up to a certain cutoff date cannot know what happened after that date. No amount of prompting can change this: the information simply is not there. Second, hallucinations are hard to control. When a model does not know something, it tends to confabulate plausible-sounding answers rather than admitting ignorance. This is not a bug that can be patched; it is an emergent property of how these models are trained. Third, multi-fact reasoning is unreliable. Even when a model has all the relevant facts in its parameters, chaining them together in a single inference step is error-prone. The model may correctly know that Radium is used in cancer treatment, and separately that Marie Curie discovered Radium, but fail to connect these facts reliably when answering a question that requires both.

Because of these limitations, most modern LLM systems rely on external knowledge sources that can be updated, verified, and queried with precision. Two architectural patterns dominate this space: knowledge bases and knowledge graphs. Although these terms are often used interchangeably, they represent fundamentally different ways of organizing information, and those differences have significant implications for what kinds of questions an LLM system can answer reliably.

Kubernetes 1.36: What's New — GA Features, Removals & Upgrade Guide

sait.butun@kloia.com (Sait Bütün) — Fri, 10 Apr 2026 08:33:28 GMT

While Kubernetes started its journey as an orchestration platform for web services, it has since transformed into something much broader. With the rise of AI, it is no longer just about orchestration; it is becoming a factory where AI workloads can live and run alongside traditional services. Each release nudges it further in that direction.

Enterprise GPU-as-a-Service Architecture with Red Hat OpenShift AI

emre.kasgur@kloia.com (Emre Kasgur) — Tue, 24 Mar 2026 09:49:27 GMT

The demand for GPU compute in enterprise environments has exploded.

Amazon Bedrock Claude ValidationException: cause and solution

Ata Ağrı — Sun, 15 Mar 2026 14:25:47 GMT

If you recently tried to invoke an Anthropic Claude model through Amazon Bedrock and hit a cryptic ValidationException, you are not alone. A policy change introduced in October 2025 significantly altered who can access Claude models via Bedrock and many developers and teams have been caught off guard. Here is what happened and what you can do about it.

OpenShift vs Harvester: Choosing a VMware Alternative

omer.urhan@kloia.com (Omer Faruk Urhan) — Tue, 10 Feb 2026 14:29:58 GMT

As VMware licensing costs continue to rise, many infrastructure teams are rethinking their virtualization strategy. The question we hear more often now isn’t “Should we modernize?” but rather “What should we modernize to?”

Kubernetes-native platforms are quickly becoming the default answer. Instead of running separate stacks for virtual machines and containers, organizations are looking for solutions that can handle both, ideally under a single control plane.

Two names that frequently come up in this conversation are SUSE Harvester and Red Hat OpenShift.

Both promise modern, Kubernetes-based virtualization. Both can replace traditional hypervisors in on-prem environments. But once you look closer, their design philosophies, and the experience they offer, are quite different.

Here’s what stands out after evaluating both platforms in real-world scenarios.

AWS GenAI Transform SQL Performance & Use Cases | Kloia Blog

tarik@kloia.com (Tarık Guven) — Thu, 08 Jan 2026 20:45:18 GMT

After years of manually wrestling with MSSQL to PostgreSQL conversions using SCT and DMS, I finally got my hands on AWS Transform SQL. I put it through its paces on a production-scale database with 1,500+ stored procedures. The results? Surprisingly impressive - and a few gotchas worth knowing about.

KubeVirt on Amazon EKS for Containers and VMs | Kubernetes

ahmet.aydin@kloia.com (Ahmet Aydın) — Thu, 25 Dec 2025 10:00:00 GMT

Kubernetes is the de facto standard for container orchestration, enabling developers to easily deploy, scale, and manage containerized applications. However, there are scenarios where running virtual machines (VMs) alongside containers is necessary. This is where KubeVirt comes into play.

KubeVirt is a Kubernetes extension that allows you to manage VMs alongside containers within the same Kubernetes cluster. This blog post will guide you through the process of installing KubeVirt on an Amazon Elastic Kubernetes Service (EKS) cluster, complete with examples and best practices.

Java Virtual Threads Benchmark & Performance Analysis | Kloia

Baran Gayretli — Fri, 19 Dec 2025 12:00:00 GMT

A groundbreaking innovation called Java Virtual Threads (Project Loom) was designed to improve the concurrency mechanism and boost Java application performance. Since virtual threads are more lightweight than traditional threads, a larger number of concurrent threads can be managed more effectively.

Werner Vogels' Final re:Invent Keynote: The Renaissance Developer

derya@kloia.com (Derya (Dorian) Sezen) — Mon, 08 Dec 2025 02:58:15 GMT

Werner’s keynotes have always been inspiring rather than sales&marketing, which is why his sessions are consistently the most anticipated at every re:Invent. Attendees usually known to return back homes after his keynote, even if there’s still another day of the conference left.

Kubernetes 1.34: Fine-Grained Container Restart Policies

raja.haikal@kloia.com (Raja Haikal) — Tue, 30 Sep 2025 14:53:02 GMT

Every Kubernetes user has faced the difficulty of managing Pods with multiple containers. Modern Pods often have a main application container, a sidecar container for extra tasks (like monitoring or managing secrets), and an init container for a one-time initial process. However, all these containers were tied to the same restart rule for the whole Pod, which was often frustrating.