Credit: https://www.xebia.com/datacenter-automation
| Area | Level 0 Not started |
Level 1 Beginner |
Level 2 Basic |
Level 3 Intermediate |
Level 4 Advanced |
Level 5 Expert |
|---|---|---|---|---|---|---|
| Security and Compliance | No formal process for security around compliance. | Periodic patch rounds of OS packages. | No shared accounts. Policies defined. Up to date components. | Full audit trail; policies periodically validated. | Fully automated security and compliance checks. | Full insight into and control of security and compliance risks. |
| Operational Management | Order new hardware after systems run out of capacity. | Order new hardware when more capacity requested; hardware is project specific. | Hardware capacity shared among projects. | Periodic checks of available and used capacity. | Cost allocated per use; detection of under-used components. | Fully automated insight into capacity, cost, business value and operational burden. |
| Resiliency | Multiple single points of failure; repaired manually. | Backup/restore procedures validated periodically. | Critical components are redundant. | All components are redundant. Backup/restore procedures are automated and validated frequently. | Auto-healing. | Planned and unplanned, automated resiliency test; predictive autoscaling. |
| Observability | No monitoring and logging; end users notice downtime first. | Logging available via Ops. Basic monitoring for platform components. | Monitoring available via dashboards. Alerts go to centralized Ops team. | Logging available via self-service for every product team. | Configurable alerts and custom dashboards for product teams. | Custom metrics. Product teams configure alerts based on custom events. |
| Provisioning | No procedures, no automation; all changes performed manually and according to operator's personal insight. | Standardized procedures; basic scripts. | Provisioning using standard configuration management tools. | No manual changes; full configuration management. | Immutable infrastructure. | Infrastructure as code. |
| Data Center Delivery | No tools; no documentation. | Standardized procedures; people trained to execute. | Changes lab-tested before applied. | Features delivered in small iterations. | Automated tests for most components. | Continuous delivery of all infrastructure components. |
| Production Team Autonomy | No formal process; make request through Ops and hope for eventual delivery. | Standardized request forms picked up whenever Ops get around to it. | Standardized requests; completed in under two weeks. | Partial self-service for basic features. | Self-service portal with graphical UI. | Full self-service with graphical UI and APIs. |