Past one DevOps engineer, short of a platform team
The person who set up the clusters now also answers every pipeline question, reviews every Terraform change and is paged for all of it. Nobody owns the direction: upgrades happen when something forces them, cost gets attention when the bill surprises someone, and each team ships its own way. Hiring a head of platform means a long search for a role nobody has defined yet — and the work does not wait for it.
What the work involves
- A named platform lead from our team who owns the roadmap and the decisions behind it, with the rest of the team doing the work — the direction has an owner, and the throughput does not rest on one person.
- Cost as a standing responsibility: a reconciled baseline in the first month, then a monthly review against it, with each saving claimed once and traced to a bill line.
- A roadmap in quarters, kept in your repository: what changes, in what order, what each step is worth and what it depends on, with the options we rejected written down.
- Foundations before features: GitOps delivery, observability and a security baseline, so every later change is visible, reversible and reviewed. the security baseline, item by item →
- Engineering routine treated as part of the platform: an upgrade policy, an incident and post-mortem habit, a golden path for new services, and lead time and change failure rate measured from your own pipeline rather than estimated.
- Architecture and vendor decisions reviewed before they are made rather than after — a second pair of eyes on the proposal your team is about to commit to.
- When you are ready to hire your own platform lead: the role written from the work rather than from a template, help with the technical interviews, and a hand-over with an agreed overlap.
How the work goes
Month 1 · Inventory and cost
Read-only access and fourteen days of measurement: what runs, what it costs, who owns each line of the bill, and where the platform has no headroom left. It ends in the FinOps audit report, yours to keep.
Month 2 · Target platform and roadmap
Where the platform should be in a year and the order of getting there, agreed with whoever owns engineering priorities. Every decision is recorded with the options rejected.
Month 3 · Foundations
The roadmap's first items land as pull requests your team reviews — usually GitOps delivery, observability and the security baseline — so everything that follows is visible and reversible.
From month 4 · Lead and run
The platform operated day to day, cost reviewed monthly against the month-one baseline, and the roadmap moved forward and re-planned each quarter, with a written summary every month.
During ongoing operation
- Who leads
- A senior engineer from our team is named as your platform lead and owns the roadmap. Who that is can change — someone is ill, someone leaves — which is why the roadmap, the decision records and the runbooks live in your repository, so whoever takes over starts from what you can read too.
- The monthly review
- An hour with whoever owns engineering priorities. The written summary arrives first, so the hour goes on decisions rather than on status.
- Day-to-day operation
- On the same terms as Kubernetes management: upgrades, capacity, and alerts answered inside the stated window. the terms →
- Cost
- Reviewed every month against the baseline from month one, measured by the MetrixForge agent where you keep it, or by the billing export where you do not.
- When a change is not agreed
- Anything that moves scope, spend or the roadmap comes to you as a proposal first. Urgent actions during an incident follow only a runbook agreed before, and are written up afterwards.
- The monthly summary
- Changes, incidents, cost and where the roadmap stands. sample →
Is this the right call?
A good fit when
- SaaS teams that have outgrown one DevOps engineer doing everything and are not ready to fund a platform department
- A CTO who has become the platform owner by default and wants the time back
- Teams about to hire their first platform lead who want the role defined by the work first
Not us when
- An interim CTO or line management of your engineers — we lead the platform, not the people
- A title on the org chart with no authority to change anything — a roadmap nobody may act on is a document, and we would rather say so on the first call
- Estates we are not allowed to observe — we do not lead what we cannot measure
What we need, and what we will never ask for
What we need
- A read-only role in the billing account.
- Our agent with read-only RBAC and no privileged containers, where you keep it.
- A branch or fork we may open pull requests from.
- An hour a month with whoever owns engineering priorities — a CTO or head of engineering — and whoever reviews pull requests in between.
What we never ask for
- Administrator credentials in the cloud account or the cluster.
- Merge rights on your repositories.
- Your source code.
- A migration onto a platform of ours.
Questions
How is this different from Kubernetes management?
Kubernetes management runs the clusters you have. This adds ownership of where the platform goes next — the roadmap, the cost review, the decisions and the engineering routine around them — with the clusters run on the same terms.
How is it priced?
The first three months are a fixed scope, priced before start. From month four it is month by month against a written scope, with a summary each month. Either side can end it with 30 days' written notice.
What happens when we hire our own platform lead?
That is a good outcome, and the engagement is built for it. The roadmap, the decision records and the runbooks are already in your repository; we agree an overlap, hand over, and either step back or stay on for day-to-day operation.