Training Agenda

Containerd

Containerd is the industry-standard container runtime that runs under Kubernetes, replacing Docker as the cluster-level container engine after Kubernetes deprecated the Docker shim. Understanding containerd matters for anyone debugging container startup, image pulls, runtime security, and low-level cluster behavior. This training covers containerd architecture, CLI tooling, image management, snapshotter configuration, and integration with Kubernetes.

1 day On-site, remote, or hybrid Up to 20 participants German or English
What We Cover
The runtime layer beneath every Kubernetes cluster
Day 1

Architecture, CLI & Kubernetes Integration

  • Containerd architecture: containerd daemon, containerd-shim, runc, snapshotters — the full runtime stack
  • CRI plugin: How Kubernetes talks to containerd via the Container Runtime Interface — the integration point
  • nerdctl: The Docker-compatible CLI for containerd — daily operations without Docker
  • ctr: Low-level containerd CLI for debugging — direct interaction with the daemon
  • Image management: Pull, push, tag, convert — OCI and Docker image formats in containerd
  • Namespace isolation in containerd: k8s.io vs default namespace — understanding the isolation model
  • Snapshotters: overlayfs, native, devmapper, zfs — choosing and configuring the right snapshotter
  • Runtime configuration: Sandbox images, registry mirrors, insecure registries — customizing the runtime
  • Content store and lease management: How containerd stores and garbage collects image data
  • Debugging container startup: containerd logs, shim state, runc state — systematic startup debugging
  • Integration with Kubernetes: crictl, sandbox inspection, image garbage collection — node-level operations
Learning Outcomes
What your team walks away with

A clear mental model of the container runtime layer and the tools to debug it — from image pull failures to container startup issues at the containerd level.

Book the Containerd training

Ideal for Kubernetes administrators and SREs who need to understand and debug the runtime layer beneath their clusters.

Get in touch