About these case studies. They describe hands-on engineering work carried out by the engineer behind MK Digital Advisory. Company names are withheld for confidentiality. Figures are stated only where they were measured; where no number is given, none is claimed.

Multi-cluster Kubernetes infrastructure on AWS and DigitalOcean

A SaaS company running 30–35 services relied on a manual release process that took one to two hours per deployment.

What was done

  • Kubernetes clusters provisioned on AWS and DigitalOcean with Terraform
  • Separate development and production environments built from the same infrastructure code
  • 30–35 services migrated onto the clusters
  • GitOps deployments introduced, with Bitbucket Pipelines building and Argo CD deploying

Result

1–2 hours → 10–15 minutes

Deployment time across 30–35 services, with an automated release process the whole team can use.

Technology

  • AWS
  • DigitalOcean
  • Kubernetes
  • Terraform
  • Bitbucket Pipelines
  • Argo CD

CI/CD security and cost optimisation across 30+ repositories

Pipelines across more than 30 repositories needed a consistent review of how securely and how efficiently they built and deployed software.

What was done

  • Pipeline configuration audited in 30+ repositories
  • Pipeline security improvements implemented from the audit findings
  • Build and deployment costs reviewed and optimised

Result

Security improvements and cost optimisations were applied to the audited pipelines. No percentage figures are published for this project.

Technology

  • CI/CD pipelines
  • Pipeline security
  • Build cost optimisation

Production observability with Prometheus and Grafana

An engineering team of up to 50 people needed a shared, reliable view of production health and earlier warning of incidents.

What was done

  • Metrics collection with Prometheus and dashboards in Grafana
  • Custom health metrics added for the signals that mattered to the product
  • Alerting configured for incident detection
  • Centralised logging set up alongside monitoring

Result

Monitoring, alerting and centralised logging in production use by a team of up to 50 engineers.

Technology

  • Prometheus
  • Grafana
  • Custom health metrics
  • Alerting
  • Centralised logging

Infrastructure security modernisation

Access to internal systems and the handling of secrets and certificates needed to move to a model with fewer standing privileges and less public exposure.

What was done

  • Network access moved to a VPN-first model following a zero-trust approach
  • Okta single sign-on rolled out across roughly 45 internal and external systems
  • Secrets management moved to HashiCorp Vault
  • TLS/PKI certificates migrated

Result

VPN-first access, single sign-on across about 45 systems, Vault-managed secrets and migrated TLS/PKI certificates in place.

Technology

  • VPN
  • Zero trust
  • Okta SSO
  • HashiCorp Vault
  • TLS / PKI