Plavi Pas contact ↓
plavi pas · devops consulting

Your project
companion.

We make your infrastructure reliable, available and secure, and bring DevOps practices that let your team ship faster with fewer incidents. We stay next to your project, not above it.

ls ~/services

What we take on.

Your stack, your budget. From a first cluster to a platform serving hundreds of thousands of requests per second.

/devopsProcesses that let your team ship often and safely: release flow, environments, on-call and incident routine, postmortems that change things.
/reliabilityInfrastructure that stays up and recovers on its own: high availability, backups that actually restore, zero-downtime deploys, load testing before it matters.
/securityZero Trust by default: no implicit trust inside the network, every person and service proves who it is, access is least-privilege and short-lived. Plus secrets management, network policies and audit trails. Hands-on SOC 2 work.
/costRight-sizing and cleanup. The cloud bill goes down, reliability doesn't.
/observabilityMetrics, logs and alerts that wake the right person, not everyone.
/cloudAWS, Google Cloud and on-premise. Multi-region, high availability, migrations without a weekend of downtime.
/kubernetesMoving to Kubernetes, or making the one you have boring again. Autoscaling that follows real load.
/gitopsArgo CD keeps every cluster in sync with Git: each change is reviewed, each state reproducible, a rollback is just a revert. CI/CD that people actually trust.

→ resultA reliable, efficient system that scales easily and is fully observable, built on proven best practices.

$ history

Been there, fixed that.

More than ten years of production infrastructure, most of it high-load. A few things we have done:

250K rpsMoved a high-load platform from bare metal to Google Cloud and from Docker to Kubernetes, with the business running the whole time.
multi-clusterHighly available Kubernetes on bare metal and multi-cluster GKE, all of it managed through GitOps.
zero downtimeDeploys nobody notices: graceful shutdown, Gateway API, load-testing farms before every big release.
scalabilityInfrastructure built so nothing stops it from growing or shrinking: scale up for a launch, scale down after it, automatically on real traffic or with one change in Git.
everything as codeTerraform and Terragrunt across AWS and GCP, Helm, Argo CD, Flux. Nothing clicked by hand.
audits passedSOC 2 requirements implemented in production, not on slides.
the extra bit

Plus our own AI tooling.

Engineering comes first. On top of it we run our own AI-agent harness, and our engineers stay in the loop at every step: we design how each step works, check what it produced and decide what moves on. After release, agents keep watching production with us.

to production
  1. brief
  2. plan
  3. build
  4. qa
  5. second opinion
  6. security
  7. merge
in production
  1. deploy
  2. watch metrics
  3. rollback ready
  4. monitoring + AI hints
  5. security scans + AI hints
human in the loopour expertise at every step: setting it up, judging the result, approving the next one
deployGitOps sync to the cluster, progressive where it matters.
watch + rollbackError rates, latency and saturation right after the release; one step back if they move the wrong way.
monitoring + AI hintsClassic monitoring: metrics, logs and alerts. On top of it AI hints: when something fires, an agent suggests the likely cause and where to look first.
security scans + AI hintsRegular security scanning of configs, access and dependencies. AI adds hints on top: which findings matter for this setup and how to fix them. An engineer decides.
handover

Then your team takes over.

We build the platform

Infrastructure, pipelines, monitoring and the processes around them, documented and in Git.

Your team runs it

Your developers take over the day-to-day changes, with our docs, runbooks and, if they like, the same AI-assisted workflow. We stay around for architecture, reviews and the weird stuff.

hourlymonthly subscriptionremote · Europe