Uploaded April 2026 | Updated September 2026, 1 day ago
In this lively AMA session, Viktor and Scott tackle a wide range of Kubernetes and DevOps questions, kicking things off with a friendly debate about Flux vs. Argo CD. They discuss secret rotation using External Secrets Operator (ESO) and the Stakater Reloader, building container images in GitHub Actions without a Docker daemon (covering Kaniko, Buildpacks, and Podman), and tools like KubePug for tracking Kubernetes API deprecations across versions. The conversation also covers bare metal Kubernetes setups with Cluster API, Metal3, Canonical MAAS, K3S, KubeVIP, and MetalLB, as well as strategies for zero-downtime deployments using Pod Disruption Budgets.
Viktor and Scott dive deep into the Flux vs. Argo CD debate, breaking down their fundamental differences — from Kubernetes-native design philosophy and target audience to how each tool handles Helm chart deployments. They share thoughts on building your own cloud using tools like VMware Cloud Foundation, Crossplane, and KubeVirt, and discuss the role of AI in Kubernetes workflows, including inference tooling from KubeCon Amsterdam and agentic debugging with observability stacks. The duo also covers DevOps fundamentals (Linux, networking, and programming), the value of home labs and certifications, Helm chart management for internal vs. external use, and best practices for managing Kubernetes resource APIs with tools like Kro and Crossplane.
▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬
00:00 Intro (skip to first question)
06:01 Secret rotation with ESO and zero downtime rollouts
07:34 Building container images in GitHub Actions without Docker daemon
12:03 Cluster API with Metal3 for bare metal Kubernetes
13:52 CNCF tools for tracking Kubernetes API changes across versions
18:02 Integrating AI agents in CI/CD for automatic code review
19:44 What stack would you choose to build your own cloud
24:43 Advice for OpenShift admin wanting to learn full DevOps
29:22 Disposable bare metal Kubernetes distro recommendations with load balancer
31:59 Achieving zero downtime during Kubernetes node group updates
33:34 Is building a production-grade home lab worth it today
36:25 Tekton vs GitHub Actions vs Azure Pipelines for CI
39:31 KubeCon Amsterdam trends and running AI on Kubernetes
44:18 Managing Kubernetes resource lifecycle with CRDs vs Crossplane
49:08 How DevOps and SREs should integrate AI agents into workflows
54:11 Helm charts of Flux Helm releases - opinions and issues
58:22 Differences between Argo CD and Flux CD explained
01:08:25 What applications are people building using DevOps
01:08:47 Are Kubernetes certifications like CKAD still worth it
01:10:31 Mirroring public Helm charts and retagging images to private registry
01:13:48 Recommended patterns for cross-provider API CRD design and CR limits
In this lively AMA session, Viktor and Scott tackle a wide range of Kubernetes and DevOps questions, kicking things off with a friendly debate about Flux vs. Argo CD. They discuss secret rotation using External Secrets Operator (ESO) and the Stakater Reloader, building container images in GitHub Actions without a Docker daemon (covering Kaniko, Buildpacks, and Podman), and tools like KubePug for tracking Kubernetes API deprecations across versions. The conversation also covers bare metal Kubernetes setups with Cluster API, Metal3, Canonical MAAS, K3S, KubeVIP, and MetalLB, as well as strategies for zero-downtime deployments using Pod Disruption Budgets.
Viktor and Scott dive deep into the Flux vs. Argo CD debate, breaking down their fundamental differences — from Kubernetes-native design philosophy and target audience to how each tool handles Helm chart deployments. They share thoughts on building your own cloud using tools like VMware Cloud Foundation, Crossplane, and KubeVirt, and discuss the role of AI in Kubernetes workflows, including inference tooling from KubeCon Amsterdam and agentic debugging with observability stacks. The duo also covers DevOps fundamentals (Linux, networking, and programming), the value of home labs and certifications, Helm chart management for internal vs. external use, and best practices for managing Kubernetes resource APIs with tools like Kro and Crossplane.
▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬
00:00 Intro (skip to first question)
06:01 Secret rotation with ESO and zero downtime rollouts
07:34 Building container images in GitHub Actions without Docker daemon
12:03 Cluster API with Metal3 for bare metal Kubernetes
13:52 CNCF tools for tracking Kubernetes API changes across versions
18:02 Integrating AI agents in CI/CD for automatic code review
19:44 What stack would you choose to build your own cloud
24:43 Advice for OpenShift admin wanting to learn full DevOps
29:22 Disposable bare metal Kubernetes distro recommendations with load balancer
31:59 Achieving zero downtime during Kubernetes node group updates
33:34 Is building a production-grade home lab worth it today
36:25 Tekton vs GitHub Actions vs Azure Pipelines for CI
39:31 KubeCon Amsterdam trends and running AI on Kubernetes
44:18 Managing Kubernetes resource lifecycle with CRDs vs Crossplane
49:08 How DevOps and SREs should integrate AI agents into workflows
54:11 Helm charts of Flux Helm releases - opinions and issues
58:22 Differences between Argo CD and Flux CD explained
01:08:25 What applications are people building using DevOps
01:08:47 Are Kubernetes certifications like CKAD still worth it
01:10:31 Mirroring public Helm charts and retagging images to private registry
01:13:48 Recommended patterns for cross-provider API CRD design and CR limits
