New: CentriCall AI voice agents that answer, qualify, and book around the clock
Centricone Technologies
Continuous integration & delivery

CI/CD and continuous delivery consulting

Most teams that call a CI/CD consultant do not have a pipeline problem. They have a delivery problem that the pipeline has made visible — a test suite nobody trusts, environments assembled by hand, a release that needs three people awake. We work on the whole path from commit to customer, because fixing the middle of it in isolation moves the queue rather than clearing it.

Book a delivery assessmentAll DevOps services

Where the time actually goes

A faster pipeline does not make a slow team fast

Build minutes are the cheapest thing to optimise and almost never the constraint. The constraint is usually a batch size nobody has questioned, a manual approval that exists because a change broke something in 2021, or a suite so flaky that a red build is read as noise. We start by finding which of those is actually binding, and we say so even when the answer is that your pipeline is fine.

Scope

What a CI/CD engagement covers

Delivery assessment against the four keys

Deployment frequency, lead time for changes, change failure rate, and time to restore — DORA's four keys, measured from your own version control and deployment history rather than from a survey. It takes a week and it is the only way to know whether a change helped.

Branching and batch size

Long-lived branches are the most common root cause we find and the least commonly suspected. Moving toward trunk-based development with feature flags is usually a bigger win than any pipeline change, and it is also the hardest sell, so we make the case with your own merge data.

Test strategy the pipeline can rely on

A gate is only a gate if a red result stops the line. That means quarantining flaky tests rather than tolerating them, putting the slow suites where they do not block feedback, and being explicit about what each stage is actually asserting.

Deployment strategy and rollback

Blue/green, canary, or rolling — chosen against your traffic shape, your data migrations, and what you can actually observe. Rollback is designed first, because a deploy you cannot reverse is a deploy nobody wants to do on a Friday.

Pipeline security and supply chain

Secrets out of the pipeline definition, least-privilege runners, dependency and image scanning as gates rather than reports, and build provenance where a customer or auditor asks for it. This is where CI/CD work and SOC 2 evidence overlap.

Handover, so the practice outlives the engagement

Runbooks, an on-call rotation your team owns, and pipeline definitions in your repository under your review process. A consultant who leaves behind a system only they understand has sold you a dependency.

Moving between CI platforms

Platform migrations are a large share of this work, and they are rarely just a translation of one YAML dialect into another. The pipeline that exists usually encodes years of undocumented decisions, and the honest first step is separating the ones that still matter from the ones nobody has questioned since.

Talk it through
  • Jenkins to GitHub Actions, GitLab CI, or Azure Pipelines — usually driven by the cost of maintaining controllers and plugins
  • Self-hosted to managed runners, with the network and secrets access that actually blocks it worked out before the cutover
  • Consolidating several teams' pipelines onto shared, versioned workflows without turning them into a platform team's bottleneck
  • Running old and new in parallel until the new one has earned trust, rather than cutting over on a date

Tooling

What we work in

Pipelines

  • GitHub Actions
  • GitLab CI
  • Azure Pipelines
  • Jenkins
  • CircleCI
  • Argo CD
  • Flux

Infrastructure and delivery

  • Terraform
  • Pulumi
  • Helm
  • Kubernetes
  • Docker
  • AWS CodeDeploy

Quality and supply chain

  • SonarQube
  • Trivy
  • Snyk
  • Playwright
  • Testcontainers
  • Sigstore

Questions

What teams ask before they start

Three things, in order. Measure how long a change currently takes to reach production and where it waits. Fix the binding constraint, which is as often batch size or test trust as it is tooling. Then hand the practice over — pipeline definitions in your repository, runbooks your team wrote, and an on-call rotation you own. If we leave and your delivery slows back down, the engagement did not work.
Continuous integration means every change merges to a shared trunk frequently and is verified automatically. Continuous delivery means every change that passes is releasable at any moment — the decision to ship is a business one, not a technical one. Continuous deployment goes one step further and ships automatically with no human gate. Most teams asking for the third want the second, and the second is where nearly all the value is.
The assessment is about a week. What follows depends entirely on what it finds: a pipeline rebuild on a single well-factored service can be a few weeks, while moving an organisation off long-lived branches is a quarter or more because the hard part is people rather than YAML. We scope the second phase after the first, not before.
We work with what you have by default. Replacing a CI platform is disruptive and is worth doing only for a specific reason — controller and plugin maintenance cost, a security model that no longer fits, or consolidation across teams. If your Jenkins works and someone maintains it, keeping it is usually the right answer and we will say so.
DORA's four keys, taken from your version control and deployment history rather than from opinion: deployment frequency, lead time for changes, change failure rate, and time to restore service. We take a baseline before touching anything, which is the part teams usually skip and then cannot prove anything afterwards.
Yes, and in practice most of this work is cloud work — the pipeline, the infrastructure it deploys, and the identity between them are one system. Where the engagement is mostly architecture and migration rather than delivery practice, the cloud solutions page is the better starting point.

Related