[CASE STUDIES]

Concrete results.
Not promises.

Every project is a real challenge solved with measurable impact. See what we've done.

01 SOFTWARE HOUSE · POLAND · 60 APPLICATIONS

APACHE
DOCKER
→
KUBERNETES
GITOPS

READ ARTICLE →
BEFORE
  • ~50% Docker, ~50% Apache2
  • Manual and semi-automated deploys
  • Rollback via SSH + backup
  • Scattered configuration
  • No environment separation
AFTER
  • RKE2: separate cluster per environment
  • GitOps with ArgoCD
  • Rollback with a single Git commit
  • Universal Helm Chart
  • Full DEV/RC/PROD repeatability
60 APPLICATIONS
MIGRATED
-87% DOCKER IMAGE
SIZE REDUCTION
~30s BUILD TIME
(FROM ~6 MINUTES)
02 SOFTWARE HOUSE · POLAND · 200 vCPU · 320 GB RAM

GITLAB
DOCKER
→
KUBERNETES
EXECUTOR

READ ARTICLE →
BEFORE
  • 5 Docker runner hosts
  • Pipeline saturates 100% CPU: 8-9x/week
  • Runner crash: ~3x/week
  • Job slowdown: daily
  • cgroups limits don't isolate
AFTER
  • Kubernetes executor (RKE2)
  • ResourceQuota per project
  • Shell-operator for per-service limits
  • 0 runner crashes/month
  • 40-60% faster pipelines
40-60% FASTER
PIPELINES
0 RUNNER CRASHES
PER MONTH
0 MANUAL
INTERVENTIONS

Have a similar challenge?
Let's talk.

DEVOPS CALL → IT CALL →