Skip to content
Book a free 30-minute call with an engineer

Kubernetes development

Kubernetes when you need it, and an honest answer when you don't

Amazon EKSGoogle GKEAzure AKSHelmArgo CDPrometheusGrafanacert-managerTerraformIstio

Cloud & DevOps

Building with Kubernetes

Kubernetes is great at running many services reliably, and it's a lot of machinery for a product with one API and a worker. We'll tell you plainly which side of that line you're on. When it is the right call, we set up managed clusters on EKS, GKE or AKS so you aren't running the control plane yourself.

Our clusters are defined in Terraform, applications ship as Helm charts, and Argo CD keeps what's running in sync with Git. Prometheus and Grafana come configured with alerts that matter, and we document the setup so your team can run it after we leave.

api-deployment.yamlYAML
1apiVersion: apps/v1​
2kind: Deployment​
3metadata: { name: api }​
4spec:​
5 replicas: 3​
6 selector:​
7 matchLabels: { app: api }​
8 template:​
9 metadata: { labels: { app: api } }​
10 spec:​
11 containers:​
12 - name: api​
13 image: registry.example.com/api:1.8.2​
14 ports: [{ containerPort: 8080 }]​
15 readinessProbe:​
16 httpGet: { path: /healthz, port: 8080 }​

The kind of Kubernetes code our team writes every week.

What we build with Kubernetes

  1. 01

    Cluster setup

    Managed clusters on EKS, GKE or AKS with networking, node pools and autoscaling configured.

  2. 02

    Helm charts

    Reusable charts for your services with per-environment values and sensible resource limits.

  3. 03

    GitOps deployments

    Argo CD watches your repository and rolls out changes, with an audit trail of every release.

  4. 04

    Monitoring and alerting

    Prometheus, Grafana and log aggregation with alerts tuned so they don't get ignored.

  5. 05

    Cluster reviews

    We audit existing clusters for cost, security and reliability, then fix what we find.

  6. 06

    Moving to Kubernetes

    Migrating services from VMs or simpler container platforms once the scale justifies it.

Why teams pick Kubernetes

01

Self-healing workloads

Crashed containers restart automatically and traffic moves away from unhealthy nodes.

02

Scales per service

Each service scales on its own metrics, so you pay for capacity where it's actually needed.

03

Same platform on any cloud

Workloads move between AWS, GCP and Azure with far less rework than cloud-specific services.

04

Safe rollouts

Rolling and canary deploys with automatic rollback when health checks fail.

Good fits

  • Microservice-based SaaS platforms
  • High-traffic APIs
  • Multi-tenant workloads
  • ML model serving
  • Hybrid and multi-cloud deployments

Where we use it

Kubernetes questions

Do we actually need Kubernetes?

Maybe not. If you run a handful of services, Cloud Run, ECS Fargate or Azure Container Apps will cost less and need less upkeep. Kubernetes earns its keep with many services, strict isolation needs or a platform team to own it.

Which managed Kubernetes service do you recommend?

Usually the one on the cloud you already use. GKE Autopilot is the most hands-off, EKS fits teams deep in AWS, and AKS makes sense alongside Entra ID and other Azure services.

How do you deploy to Kubernetes?

With GitOps. Your CI builds and tests images, updates a version in Git, and Argo CD applies the change to the cluster. Every deploy is reviewable and reversible.

Can you take over and improve our existing cluster?

Yes. We start with an audit of security, cost, resource limits and monitoring, then work through a prioritized list of fixes without interrupting production.

Pairs well with