DE EN
Contact
Cloud engineering

DevOps and CI/CD: deployments that become routine

When a release frays nerves, the cause is rarely the code and usually the road to production. Sortia Labs automates build, tests and deployment of your applications with GitHub Actions, GitLab CI or ArgoCD and keeps the infrastructure as code in git. Every change becomes traceable, and a rollback is planned instead of improvised.

Request a first call Cloud engineering
  • DevOps consulting
  • CI/CD pipeline
  • GitHub Actions
  • GitLab CI
  • Infrastructure as code
  • GitOps
DevOps and CI/CD

What we put in place in DevOps projects

CI/CD pipelines

Every push builds the application, runs the tests and produces an artefact that goes unchanged to test and production. We set up pipelines with GitHub Actions or GitLab CI, matching your repository.

Infrastructure as code

We describe servers, networks, DNS and cloud resources with Terraform and Ansible. Changes go through the same review as application code, and an environment can be rebuilt from the repository.

Automated checks

Unit and integration tests, tests against a real database and a check of dependencies against known security vulnerabilities. If a check fails, the pipeline stops before the version reaches a server.

Staging and rollback

New versions land on a staging environment first and go to production after approval. The previous version stays ready, and the way back is documented and rehearsed.

Secrets and access

Passwords, API keys and certificates do not belong in the repository. We keep them in a secret store or in protected pipeline variables, readable only by the steps that need them.

Making DevOps part of the team

Tools alone change little. We work with development and operations on shared routines, short reviews and clear responsibilities, so that your team can maintain the pipeline itself.

Approach

How we introduce DevOps in your company

  1. 1

    Current state

    We walk through a real release with your team: who builds, who tests, who deploys, and where does work wait? The bottlenecks show where automation helps first.

  2. 2

    First pipeline

    One application gets a complete pipeline from commit to the staging environment. Your team uses it day to day, and its feedback shapes the pattern before we roll it out further.

  3. 3

    Expansion

    More applications, infrastructure as code, deployments to production and rollback. Every stage comes with documentation so that your team can run it on its own.

  4. 4

    Maintenance

    Pipelines age along with the code. We keep runners, dependencies and tools up to date or hand maintenance over to your team with a briefing and a runbook.

From our own work

Software we build and run ourselves.

Postkey Encrypted vault and sharingWe ship Postkey by script from git to our server in Germany: build, upload, run database migrations, restart the services and then check that the application answers from the internet. Tempo Calendar and day plannerFor Tempo, the configuration and services of the server live in git and reach the server by script. The secrets stay on the server and are kept out of the repository.

Questions about DevOps and CI/CD

What is the difference between CI and CD?

Continuous integration means that every change is built and tested automatically. Continuous delivery goes further: the tested version is brought automatically up to the point of release and goes live after approval. With continuous deployment even that click disappears. Which stage fits depends on your approval rules.

GitHub Actions or GitLab CI: which fits?

Usually it depends on where your code already lives. With GitHub, GitHub Actions is the natural choice; with GitLab, its built-in CI. For Kubernetes we often add ArgoCD, which keeps the cluster in line with the repository. Switching code platforms just for the pipeline rarely pays off.

Is DevOps worth it for small teams?

Especially there. When two or three developers also run operations, a pipeline saves manual work on every release, and the knowledge no longer depends on a single person. We start small: one pipeline, one test system, one documented way back.

Does CI/CD work with older applications too?

Yes, and it often pays off there in particular. Even an older PHP or Java application can be built, tested and deployed automatically. Where tests are missing, we start with checks of the core workflows and extend them with every change.

Related topics

Read on.

KubernetesKubernetes plays to its strengths when many services, several teams or strongly fluctuating load come together.MonitoringA service that stops at night is often noticed only when the first customer calls.Cloud migrationMoving to the cloud, to another provider or back to your own servers does not have to bring your business to a halt.Web app developmentSortia Labs builds web applications that your staff and customers use in the browser: customer portals, internal tools and admin interfaces.Legacy modernizationThe application has run for years, but hardly anyone dares to touch the code any more: the original developer is gone, and the PHP version or the…AWS and AzureAWS and Azure offer an almost endless choice of services, and that is exactly what makes getting started hard.

All about Cloud engineering All services and topics

How does your code reach production today?

Describe your current process in a few sentences. We will tell you which step can be automated first and what it brings your team.

Request a first callCloud engineering