Cloud Architecture
AWS · Azure · GCP · DigitalOcean · Hetzner · Contabo
Cloud bills grow faster than workloads. Instances get sized for a launch-day guess and never revisited, snapshots accumulate, managed services get bought for things a single tuned server already does well, and nobody can say which resources are load-bearing. The opposite failure is just as common: one oversubscribed box running database, application and backups together, with no plan for the day it dies.
- Inventory of current resources and spend across every provider account, including the orphaned volumes and snapshots that never show up in planning
- Workload profiling: real CPU, memory, IOPS and network use over a normal week, not the peak used to justify the original sizing
- Target architecture sized to that profile, with the separation that matters, application, database and backup storage, and a stated failure mode for each component
- Security and access review: firewall rules, SSH access, IAM roles and secrets handling
- Migration or right-sizing plan with expected monthly cost and a rollback path
- Disaster recovery plan with a stated recovery point and recovery time objective, then a real restore rehearsal
From $200 (infrastructure assessment report)
See the full rate card for fixed packages and managed retainer plans.
Deliverables
- Cloud infrastructure design and deployment
- Multi-cloud and hybrid architecture
- Auto-scaling and load balancing
- Security hardening and IAM configuration
- Cost audit and optimization
- Disaster recovery planning
Which cloud provider do you recommend?
It depends on the workload, and we have production experience across AWS, Azure, DigitalOcean, Hetzner and Contabo. For a single-region Odoo instance a good VPS is usually the cheapest sound answer; the hyperscalers earn their premium when you genuinely need their managed services or regional footprint.
Can you cut our cloud bill without downtime?
Often yes. Deleting orphaned volumes and snapshots, right-sizing over-provisioned instances during a maintenance window, and dropping managed services that duplicate what your server already does are the usual wins.
Do you do multi-cloud?
Where there is a reason for it, such as keeping backups in a different provider from production. We do not recommend multi-cloud as a default, because it doubles the operational surface for most teams.
What does the assessment report include?
A written document with current state, findings ranked by risk, a target architecture, cost comparison and a prioritised remediation plan you can execute with or without us.