Running a Well-Architected Review & Next Steps
A Well-Architected review is a structured conversation about a specific workload, not a one-off compliance exercise. It works best as a recurring practice.
Suggested process
- Scope the review to a single workload or system with a clear owner.
- Work through each pillar’s checklist with the people who actually operate the workload.
- Flag every “no” or “unsure” answer as a candidate finding. Don’t pre-filter.
- Prioritize findings by risk and effort. Not everything needs to be fixed immediately.
- Record findings and improvement actions in your normal backlog, with owners and target dates.
- Re-run the review after major architecture changes, and at least quarterly for production workloads.
Cadence
For new workloads, run a lightweight review before going live, focused on Security and Reliability.
For production workloads, conduct a full review on a quarterly basis, and after any major change (new region, new compliance requirement, significant scale change).
For regulated workloads, align your review cadence with your compliance/audit calendar (e.g., ahead of NIS2 or DORA-related assessments).
Getting help
Cleura’s architects and consultants can facilitate a Well-Architected review directly, particularly useful for compliance-sensitive workloads, migrations, or when validating a new architecture before launch.
Quick Reference: Pillars and Cleura Cloud Capabilities
| Pillar | Relevant Cleura Cloud capability |
|---|---|
| Operational Excellence | CCMP, CLI & API, Infrastructure-as-Code, projects & quotas, Gardener CI-CD integration, Managed Services |
| Security & Digital Sovereignty | Compliant Cloud, security groups, project isolation, EU-only regions, ISO 27001 / MSB Protection Class 3 facilities, VPN connectivity |
| Reliability | Availability zones, Load Balancer + auto-scaling, automated daily snapshots (10-day retention), multi-region DR, Managed Database/Kubernetes HA |
| Performance Efficiency | VM flavor profiles (Generic / Low Latency Disk / High Intensity CPU), block & object storage tiers, regional placement, Gardener worker pools |
| Cost Optimization | Per-second billing, storage lifecycle policies, project quotas, transparent pricing & volume discounts, Managed Services trade-off |
Next Steps
- Pick one production workload and run it through the five checklists in this document.
- Log every gap found as a backlog item, prioritized by risk.
- Confirm the workload’s deployment model was chosen deliberately based on its regulatory, security, and availability requirements. Do not choose by default.
- Schedule the next review. Quarterly is a reasonable default for production workloads.