Containers with Kubernetes

Cloud-native platforms on EKS

We build production Kubernetes platforms on AWS using Amazon EKS and CNCF tooling, for greenfield deployments and for migrations off legacy setups. Mostly for enterprises, though the same building blocks scale down without being rebuilt.

How we work

Four steps. We discover what you already run and what constrains it, design the target platform, build it as code, then enable your team to operate it.

What we build

• Managed Amazon EKS clusters, multi-AZ and production-ready

• Node lifecycle automation with Karpenter and EKS Auto Mode

• eBPF networking with Cilium, Hubble and Tetragon for observability and runtime security

• Ingress without load balancers using Cilium, or AWS-native options where they fit better

• Container image builds through GitHub Actions, GitLab CI or AWS CodeBuild

• Policy-as-code with Kyverno for cluster governance

The eBPF layer is where most of the security sits. Container and network security goes into it properly.

Container practices

Image hardening, minimal base images, and automated scanning and patching. Most container security problems start with what went into the image rather than with the cluster running it.

GitOps and automation first

Every environment we deliver is automated, reproducible and observable.

• Infrastructure as Code with OpenTofu, Helm and Crossplane

• Git as the source of truth, with declarative cluster state reconciled by Flux

• Secrets held in AWS Secrets Manager or HashiCorp Vault and pulled in rather than copied

• Shift-left security, with policy checks on pull requests and enforcement in the pipeline through Kyverno

The result is consistent operations, fast recovery and a change history nobody has to maintain by hand. GitOps in practice covers how Flux is set up.

Docker and Kubernetes Bootcamp

If your team is new to containers, start here rather than with a platform. Two days on site, building in your own AWS account. Day one covers Docker and standing up a Docker environment. Day two covers the Kubernetes components, building a cluster, and deploying a three-tier application.

You keep the environment for two weeks afterwards, with remote support while your team practises on it. Aimed at developers, architects and operations engineers, with a small group size so everyone builds rather than watches. The two days can also be booked separately.

Full agenda, group size and pricing are in the two-pager.