Cloud Server Procurement: A Checklist from Quote to Handover
A cloud server quote should identify a configuration your team can deploy and operate. A provider-wide “best” award cannot establish application compatibility, usable capacity or the support you will receive.
The earlier five-provider ranking and price table have been removed because they did not establish a like-for-like evaluation. This checklist covers the server purchase and handover rather than another global-provider scorecard.
Specify the required environment
Record the operating system, runtime, database, background jobs and access requirements. Include licence and architecture constraints before selecting a machine family.
Check whether you need a virtual server at all. A supported managed application can be a different operating choice; the configuration comparison helps separate the labels.
Get the allocation and limits in writing
| Resource | Evidence required |
|---|---|
| CPU | Shared, dedicated or burst policy and sustained limits |
| Memory | Allocation, relevant caps and failure behaviour |
| Storage | Capacity, persistence, I/O limits and recovery options |
| Network | Addressing, transfer accounting and capacity rules |
| Region | Required products, account access and current provisionability |
A vCPU is not automatically an exclusive physical core. A region catalogue entry also does not prove immediate capacity for a particular allocation.
Assign operating ownership
Identify who installs software, applies patches, monitors the application and manages access. Confirm the supplier’s boundary for custom code and database problems.
Keep billing and recovery access under organisational control. Use individual accounts and scoped service identities where supported.
Validate one representative deployment
Deploy a suitable copy and verify important tasks, logs and external integrations. Measure representative load within the supplier’s rules and record errors alongside response times.
The load-test runbook covers capacity and supported recovery checks. A successful machine boot is not acceptance of the application service.
Verify the bill and the recovery route
Record active, idle and retained-resource charges, quotas, commitments and support fees. Ask what must be removed to end billing without assuming a shutdown is enough.
Before any removal, preserve required data and verify the backup. Test restoration into a protected environment and confirm which configuration and external dependencies are included.
Record the acceptance criteria before testing so a favourable average does not conceal a failed critical task.
Complete the handover
Retain the quote, deployment instructions, monitoring routes, backup procedure and incident contacts. Ensure another authorised operator can use them without relying on one person’s memory.
Use the global region shortlist when comparing locations and the budget review for complete cost. Accept the server only when the tested configuration and remaining responsibilities are clear.








