Ship to Kubernetes daily, on a platform built one standard way

We first check whether Kubernetes suits your applications. If it does, we design one consistent platform and build it. Then we either operate it or pass it to your team.

Sound familiar?

Which of these sounds like your team?

Waiting on every release

Merged code sits for a long time before it reaches production.

No two clusters alike

Each one was configured its own way.

Is Kubernetes overkill?

It may be more platform than your applications call for.

Hard to undo a bad deploy

When a release fails, getting back to the last good version is painful.

You hear about problems from users

Complaints arrive before any alert does.

Paying for peaks

Node capacity is set for busy spells that hardly ever happen.

Process

How we work

  1. 01

    Ask whether you need it

    We study your applications and tell you plainly whether Kubernetes is worth it.

  2. 02

    Standardise the design

    Every cluster will follow a single platform design.

  3. 03

    Build

    We set up monitoring, delivery pipelines, GitOps and the clusters themselves.

  4. 04

    Operate

    Grootan runs the platform, or your team does after a handover from us.

Specifics

Technical details

  • Amazon EKS clusters whose node groups scale automatically, with each environment defined in code
  • Deployments driven from Git using Helm charts and Argo CD
  • A delivery pipeline for each service that rolls back on its own when a release fails
  • Traffic comes in through ingress and load balancers, and Istio or another service mesh is added only if a case needs one
  • Dashboards in Grafana that bring traces, metrics and logs together
  • Container images are scanned, access is role-based, secrets are managed and network policies restrict traffic

Options

Flexible to fit you

  • Run your services on KubernetesorUse lighter AWS managed services when Kubernetes would be too much
  • Start fresh on Amazon EKSorKeep and tidy up the clusters you have
  • Grootan builds the platformorWe build it together with your engineers
  • We keep running itorOwnership moves to your team

Measurement

What we measure

  • How often you deploy to production
  • Time from a commit to that change running live
  • Share of deployments that cause a failure
  • How fast service comes back after an incident

FAQ

Questions we hear

Which teams benefit from Kubernetes?

Count your services and how often they ship. If both numbers are high, Kubernetes earns its place. With a few applications, a simpler option is usually the better call.

What do we take on if we adopt Kubernetes?

Someone on your side, or ours, has to run the cluster, and that person needs specific skills. Because that cost is real, we only recommend Kubernetes after ruling out simpler options.

Our clusters are already running. Will you take them on?

Yes. You end up with clusters that follow one standard and come with documentation, and the starting point is checking how they run today.

How much does a Kubernetes project cost?

Each one differs in scope and complexity, and those two things set the price. We cover it when we first talk.

Give your developers a platform they can trust

Related

More Cloud & AWS services