Infrastructure as code
Servers you can rebuild from a repository.
- Terraform for AWS VPCs and environments
- Nomad and container orchestration
- Identical dev, staging and production
- No manual deploys
SRE, DevOps and cloud
Infrastructure that stops being the emergency.
I make infrastructure less dramatic. I move platforms from manual deploys and regular outages to automated, observable systems, alongside your team rather than handing over a deck.
-
99.95%
uptime on a platform that used to go down. Monitoring, alerting and incident practice that stuck.
-
40%
shorter CI builds, from over 60 minutes to under 35, without dropping test coverage.
-
17+
years running platforms, from single servers to multi-region.
Is this for you?
- Deploys are manual and everyone is a little afraid of them.
- You find out about outages from users, not from alerts.
- The cloud bill grows faster than the product.
What I do
Reliability
Fewer pages at night.
- Metrics, logs and APM
- Alerting people actually trust
- Incident practice and reviews
- Capacity planning
CI/CD
A pipeline the team does not route around.
- Build times down, coverage kept
- Standardised pipelines for every service
- Tests that run where they should
- Zero-downtime deploys
Cloud cost
Same product, smaller bill.
- Right-sizing instances
- Cost tracking and alerts
- Architecture reviews
- Performance kept in the trade
How it works
-
Assessment
I read the infrastructure, the pipelines and the last few incidents.
-
Design
A target architecture, and the smallest first step that lowers risk.
-
Migration
Shipped in small pieces, with downtime planned and visible.
-
Handover
Runbooks, dashboards, and a team that owns it.
Formats and duration
- Advisory
- A day or two a month. Architecture, hiring, hard calls.
- Project
- Three to six months, embedded with your engineers.
- Workshop
- One to three days on IaC, CI/CD or incident practice.
Questions
Do you work inside the team?
Yes. I pair with your engineers and write the changes with them. The handover is the point, not an afterthought.
We are on virtual machines, not Kubernetes.
Fine. I have run Nomad, VMs and containers in production. The tool matters less than the practice around it.
Will this break production?
Migrations run in small steps with downtime planned in advance. The first change is always observability, so you can see what happens next.
How does it start?
A call about the platform and what hurts. If it looks useful, I send a short written plan: the first month, the cost, and what I need from you.
Other services
What is breaking?
Tell me about the platform, the deploys and the alerts that wake you up.