Case study
Eleven environments into one platform
Consolidated eleven hand-built environments into a single orchestrated Kubernetes platform with GitOps delivery, cutting the cloud bill and the incident rate at the same time.
Measured outcomes
- -38% monthly cloud spend after decommissioning
- 1 deployment path for all product teams
- -60% production incidents over the following two quarters
Details
- Client
- B2B SaaS company, 60 engineers
- Sector
- Software
- Services used
Challenge
Where it started
Each product team had built its own environment over four years. Eleven clusters and VM fleets, six ways to deploy, no shared monitoring, and a cloud bill that had doubled while traffic grew by a third. Incidents were frequent and every one needed the person who had built that particular corner.
Approach
What I did, in order
-
Measured first. Cost attribution per environment, utilization per node pool, and an incident review of the previous twelve months.
-
Designed one platform with namespaces and quotas per team, shared ingress, secrets, observability and a single deployment path.
-
Migrated team by team using GitOps, so every workload moved as a reviewed change and could be rolled back the same way.
-
Decommissioned the old environments only after two clean weeks on the new one, with the savings tracked per step.
Deliverables
What the team kept
- Platform codebase (Terraform, Helm, Argo CD) with per-team onboarding
- Hardened base images and a CI pipeline with scanning and signing
- Shared observability stack with SLOs per service
- Cost dashboard by team and a monthly review ritual the team kept running
Consolidation is mostly deletion. The platform itself was not exotic. What made it work was the order: measure, build the target, move by reviewed change, delete only when proven.
Next step
Facing something similar?
Tell me what is slowing your team down. I answer within one business day, and the first 30 minutes are on me.