Skip to content

Cloud & DevOps

A release that no longer needs a ritual

CI/CD, containerisation, environment standardisation and infrastructure as code — so that shipping is routine work rather than an event.

The problem

One person knows how to deploy

Deployment is manual, the steps live in a chat thread, staging no longer resembles production, and rollback means pulling from a backup. Configuration was edited on the server and no longer exists in the repository — the next install cannot reproduce the same result.

  • Shipping depends on one person
  • The environments have drifted apart
  • There is no rollback plan
  • Secrets live in chat threads and .env files

What we deliver

A release that no longer needs a ritual

The pipeline

Build, test, image and deploy as one repeatable chain — with no manual steps.

Containerisation

The application and its dependencies in an image, versioned and reproducibly built.

Environment parity

Dev, staging and production from one description — the difference lives in parameters only.

Secrets handling

Role-based access, secrets outside the repository, with a rotation procedure.

Rollback and release strategy

Back to the previous version in minutes, not by restoring a backup.

Observability after the release

Metrics, logs and alerts that show the problem before the user does.

Architecture

Architecture

  1. 01CommitCode and configuration in one repository
    • Git
    • Semantic versioning
  2. 02BuildA reproducible, versioned image
    • Docker
    • Registry
  3. 03TestAutomated, with no manual steps
    • GitHub Actions
    • Lint / types
    • Smoke test
  4. 04DeployStaging and production from one description
    • Dokploy
    • Compose
    • Terraform
  5. 05ObserveMetrics and logs around the release
    • Prometheus
    • Grafana
    • Loki
  6. 06RollbackThe previous version in minutes, not from a backup
    • Blue-green
    • Canary
    • Migrations
The release chain — from commit to production, keeping a way back

Capabilities

Capabilities

Every item is marked: verified production experience, or engineering capability.

Docker and containerised services

Proven

Our own production services run in containers — the monitoring, logging and internal tooling stack included.

  • Docker
  • Compose
  • Registry

CI/CD pipelines

Proven

Build, test and deploy on GitHub Actions, with environments based on Dokploy.

  • GitHub Actions
  • Dokploy
  • Git

Database migrations and backup

Proven

Schema versioning, backup rotation and a restore rehearsal into a scratch database.

  • PostgreSQL
  • Migrations
  • pg_dump

Reverse proxy and TLS

Proven
  • nginx
  • Traefik
  • Let's Encrypt

Infrastructure as code

Capability
  • Terraform
  • Ansible
  • GitOps

Kubernetes

Capability
  • Kubernetes
  • Helm
  • Ingress

Blue-green and staged rollout

Capability
  • Blue-green
  • Canary
  • Feature flags

Application observability

Capability

Metrics and logs on the same platform that carries the infrastructure monitoring.

  • Prometheus
  • Loki
  • Grafana

Process

Process

  1. 01

    Audit

    How code ships today and where the chain breaks.

  2. 02

    Standard

    An agreed build, versioning and environment description.

  3. 03

    Pipeline

    Automation for one service, end to end.

  4. 04

    Roll out

    The same template across the remaining services.

  5. 05

    Observe

    Metrics, logs and alerts around the release.

  6. 06

    Handover

    Documentation and training for the team.

Technology stack

Technology stack

Containers
DockerComposeRegistryKubernetes
CI/CD
GitHub ActionsDokployGitSemantic versioning
Infrastructure as code
TerraformAnsiblecloud-init
Runtime
nginxTraefikPostgreSQLRedis
Observability
PrometheusGrafanaLokiAlertmanager

Engagement model

Engagement model

Project

A one-off scope: audit, migration or implementation with a fixed outcome.

Retainer

Monthly engineering hours — specialist access on demand.

Co-managed

NetWizard and your in-house team together, with split responsibility.

Use cases

Use cases

Shipping is a night job

Every release is a list of manual steps. We build the pipeline, the containers and the rollback — until shipping in the middle of a working day is possible.

Staging does not resemble production

The tests pass and production still breaks. We fix the environment description so the only difference left is data and parameters.

Cloud spend is growing unchecked

We measure what actually consumes resources, remove abandoned environments and put spend on a dashboard — so the decision rests on a number.

FAQ

FAQ

Do we need Kubernetes?

Often not. For a handful of services, Docker Compose or a lightweight platform is cheaper and needs less operating. Kubernetes wins when the scale and the team can sustain it.

Do you work in our cloud or yours?

Either. Public cloud, our datacenter or your own hardware — the pipeline and the description stay the same; only the target changes.

Will you replace our developers?

No. We build the infrastructure and the pipeline so your team ships faster. If you also need product development, that is a separate service.

Tell us about your infrastructure