From a broken cluster to 99.98% uptime
How Digital Care rescued a government online-services platform from a failing OpenShift cluster and rebuilt it on Kubernetes.
The client operates online services used by government institutions. The platform ran on an OpenShift 3 cluster whose certificates had gone invalid: pods no longer started and nodes could not reach each other. The team had no safe way to deploy or recover, and public-facing services were at constant risk of going dark.
First, resuscitate: restore cluster operation quickly. Then, rebuild for good: migrate from OpenShift to Kubernetes, stabilize the infrastructure and put proper CI/CD underneath.
Rescued first. Rebuilt to last.
Emergency recovery
Issued new valid certificates and brought the failing cluster back to life in 4 hours, putting services back on solid ground before touching anything else.
A new Kubernetes home
Deployed a fresh Kubernetes cluster, moved every application onto it and destroyed the old OpenShift cluster.
Fault-tolerant storage
External storage built on GlusterFS: data survives node failures.
CI/CD on Jenkins
Automated build and deploy pipelines: applications ship to the cluster quickly, automatically and without downtime.
Monitoring and logging
Prometheus + Grafana across all services: availability and resource consumption visible at all times.