Packages

Two ways to run production Kubernetes.

Launch and Scale are the same platform at two depths. Scale contains everything in Launch, so you start where your team actually is today and move up when the business makes you.

Tier 01

Launch

Production, without the sprawl.

For a first real production cluster, or a team whose deploys are still manual.

Where every engagement starts

Managed Kubernetes cluster in your own cloud account
Worker nodes in a single availability zone
Private network, load balancer and ingress
CI/CD pipeline: validate → build → deploy
Secret scanning that fails the build on any finding
Image scanning that fails the build on critical vulnerabilities
Prometheus and Grafana with baseline dashboards

Tier 02

Scale

High availability and a real promotion path.

For production traffic that cannot wait for someone to fix it by hand, and for teams preparing for SOC 2 or ISO 27001 who need every production change on record.

Everything in Launch, plus

Worker nodes spread across multiple availability zones
Full pipeline: build → security → scan → publish → deploy
Static code, dependency and infrastructure-as-code scanning in the pipeline
Encrypted secrets committed safely to Git
GitOps promotion across dev, staging and prod, with a person approving every production release
Admission policies and network policies
Centralised log aggregation
Horizontal autoscaling
Acceptance test: we kill a node in front of you and workloads reschedule unattended

Pricing depends on your cloud, cluster size and support level. We quote it after a scoping call, never off a rate card.

Architecture

What gets built in your account.

Select a diagram to open it full size.

Launch architecture: a GitLab pipeline outside your cloud account runs lint, secret scanning, image build and image scanning, then deploys with Helm into one Kubernetes cluster in your account. The cluster sits behind a public load balancer and ingress-nginx, runs one worker node pool in a single zone, and includes Prometheus and Grafana. — A GitLab pipeline outside your cloud account. Secret scanning and image scanning gate every deploy.
One cluster in your account: a load balancer, ingress, and one worker node pool in a single zone.
Prometheus and Grafana inside the cluster, opened with kubectl port-forward.

Opens full size in a new tab.

Scale architecture: a six-stage GitLab pipeline with code, dependency, infrastructure and image scanning updates a Git manifests repository. ArgoCD syncs dev and staging automatically and prod only after a person syncs it. The cluster spreads worker nodes across zones, or fault domains on Oracle Cloud, with Kyverno, network policies, Sealed Secrets, autoscaling, Prometheus, Grafana and Loki. — A six-stage pipeline with code, dependency and infrastructure-as-code scanning.
ArgoCD promotes dev and staging from Git. Production waits until a person syncs it.
Worker nodes spread across zones (fault domains on Oracle Cloud), with policies, autoscaling and central logs.

Opens full size in a new tab.

Cloud coverage

Every tier, on every cloud we support.

Each combination below has been built and verified on live infrastructure. Provisioned, deployed to, and torn down again. None of it is derived from a reference architecture we have never run.

Cloud	Launch	Scale
Amazon Web Services EKS
Google Cloud GKE
Microsoft Azure AKS
Oracle Cloud OKE		†

† Oracle Cloud, Scale. High availability is currently delivered across fault domains within a single availability domain. That survives rack and hardware failure and rolling maintenance, but not the loss of an entire availability domain. The multi-zone resilience described under Scale applies as written on AWS, Google Cloud and Azure. If Oracle Cloud is your target and you need multi-zone redundancy, raise it on the scoping call. It changes the architecture, and we would rather say so now than in month three.

Support

How it is run is a separate decision.

The tier decides what gets built. Support decides who operates it afterwards. The two are independent. Any tier can be handed over to your team or fully operated by us, and support is quoted separately.

Deploy and hand over

We build it, document it, train your engineers on it, and step back. The platform is yours to run.

Managed

We operate it as an extension of your team: version upgrades, security patching, alert response and regular health reporting.

Premium

Extended coverage hours, tighter response targets, and a standing architecture review as your platform grows.

Straight answers

What these packages don't do.

The limits are easier to read now than to discover during an engagement. This is the list we would give you on a call anyway.

We do not publish a delivery date.

You get a committed timeline on the scoping call, measured against your cloud and your scope, rather than a number pulled from a pricing page.

Failover is not instant on every cloud.

Scale guarantees that workloads reschedule automatically when a node dies. Load-balancer failover behaviour differs between providers, and we share the measured numbers for yours during scoping.

Dynamic testing is not included, and we are not an audit firm.

Scale covers static, dependency, image and infrastructure scanning. Dynamic application security testing (DAST) is outside its scope. Scale produces change and security evidence that teams preparing for SOC 2, ISO 27001 or Cyber Essentials Plus can use, but we do not audit or issue certifications.

Tell us about your cluster.

A free 30-minute call. No slides, no pitch. We look at your actual stack and tell you honestly what we would build.

Book a scoping call or email hello@hyltra.com