02 / Run
Services

The boring parts that keep production up.

The cloud platforms underneath your software, run by people who built software on them. Monitoring, releases, incident response, capacity, cost. We don’t resell servers. We operate the systems they run.

What we run

Four areas. Production-grade.

Always in combination — observability without auto-scaling is noise; auto-scaling without budgets is a bill.

01

Cloud platforms

We build on Google Cloud by default, on Azure where the client is committed to the Microsoft stack, and on Cloudflare where edge matters. Cloud Run for stateless services; Kubernetes (GKE Autopilot) when the workload actually earns it. Most don’t.

  • Google Cloud
  • Azure
  • Cloudflare
  • Cloud Run
  • GKE
02

Migrations

On-prem to cloud, legacy hosting to managed services, monolith to services where the business case is real. We start with a cost-and-risk model — most migrations that fail are sold on the wrong one. We won’t recommend a re-architecture you don’t need.

  • Lift-and-shift
  • Re-platform
  • Re-architect
03

Reliability & operations

We instrument first, page second. Real SLOs against the user journeys that matter, alerting only on things that require a human, post-incident review on every page-out. Our on-call cover follows Nairobi business hours by default and can be extended.

  • Observability
  • SLOs
  • On-call
  • Incident response
04

Cost & FinOps

Most cloud bills are 20–40% larger than they need to be. We bring a monthly review, a tagging discipline, and the boring spreadsheet that turns commitments into savings. Improvements net of our fee, or we’re not doing it right.

  • Budgets
  • Commitment planning
  • Right-sizing
How a run engagement works

Three shapes. Pick the one that fits.

01

One-off project

A migration, a re-platforming, a Kubernetes setup.

Fixed scope, fixed price, hand-over at the end. We document everything so your team can keep going.

02

Co-operated

You have an ops team. You want senior backup.

We work alongside your engineers on releases, incidents, and architecture review. Monthly retainer.

03

Fully managed

You don’t want an ops team. You want it to work.

We operate the platform. You get a single point of contact, a monthly report, and a phone that rings when things break.

Need someone to run it?

Tell us what you’re running today and what isn’t working. We’ll come back with a plan, a number, and the names of the two engineers who would be on your account.