Skip to content
sedwis

DevOps technologies

Tooling that makes deploying boring.

The goal isn't sophistication. It's that nobody has to think about a release.

We deliberately recommend less DevOps than most firms. Kubernetes is excellent at genuine scale and significant overhead below it. Most teams need automated tests, one-command deploys and basic monitoring — not a platform team.

What we use

And why we picked each one.

01

Docker

Containerisation.

WhyMakes 'works on my machine' a solved problem. The single highest-value DevOps tool for a small team.

02

GitHub Actions

CI/CD integrated with your repository.

WhyOur default. No separate system to maintain, and the free tier covers most projects entirely.

03

Terraform

Infrastructure as code.

WhyReviewable, reproducible infrastructure. Prevents the undocumented drift that makes recovery impossible.

04

Kubernetes

Container orchestration.

WhyGenuinely excellent at scale. We recommend it far less often than we're asked to, because the operational cost is real.

05

Prometheus & Grafana

Metrics collection and dashboards.

WhyOpen-source monitoring without per-host licensing. Alerting on the metrics that predict an incident.

06

Sentry

Error tracking.

WhyUsually the first observability tool we add. Knowing what's breaking beats guessing.

07

Ansible

Configuration management.

WhyFor servers that aren't containerised, including plenty of existing client infrastructure.

How we choose

Four criteria, in this order.

01

Fit

Does it solve your actual problem, or a more interesting one?

02

Hiring pool

Could you hire for it in India if you took the project in-house?

03

Total cost

Licences, hosting and the operational time to run it — not just build cost.

04

Exit risk

How hard would it be to leave? We don't build things you can't walk away from.

FAQ

DevOps questions.

Probably not. Managed container services or plain VMs serve most businesses better and cost far less to operate. We'll recommend Kubernetes when your scale or team genuinely justifies it.

Automated tests on every merge, one-command deploy, error tracking and uptime monitoring. That combination removes most release anxiety and takes a week or two to set up.

Yes, and incrementally is usually right. We start with whatever is costing you the most time rather than proposing a wholesale replacement.