Cloud architecture and infrastructure
Designing what your application runs on, writing it down as code, and making the path from a commit to production boring.
This is usually one of two situations: something works but nobody can safely change it, or a bill has grown faster than the business. Both are architecture problems before they are cost problems.
What the work covers
- Architecture design and review for AWS and Cloudflare
- Infrastructure as code with Terraform, SST or Pulumi, so environments can be rebuilt instead of remembered
- CI/CD pipelines that deploy on merge and roll back without drama
- Database design, migrations and migration off a provider you have outgrown
- Cost review, which frequently pays for the engagement
What we build it with
- AWS: Lambda, S3, RDS, SES, Glue
- Cloudflare: Workers, Pages, R2, D1, Queues
- Terraform, SST, Pulumi
- GitHub Actions
How an engagement runs
- 01
Read what exists
We start by reading the account and the code, not by proposing a rebuild. Most infrastructure is closer to fine than it feels.
- 02
A written assessment
What is fragile, what is costing more than it should, and what we would change first. Useful on its own even if you stop there.
- 03
Change it incrementally
In pieces, each one shippable and reversible. No big-bang migration weekend.
- 04
Hand over the runbook
How it works, how to deploy it, what breaks and what to do about it. Written for whoever comes after us.
What we don't do
- Azure and Google Cloud. We would be learning on your budget.
- Kubernetes clusters, which are usually the wrong answer at the scale we work at
- Compliance certification work such as ISO or SOC 2