Azure cost model for Diskpipe evaluations

Start with the general cost-control guide. Seek an existing cluster or suitable hosts, confirm the subscription and target resources, and agree on a small single-region evaluation. Cross-region evaluation and public internet storage traffic require explicit approval after reviewing estimated costs.

Cluster and storage

AKS offers different cluster-management pricing tiers; underlying infrastructure must also be included in the estimate. Identify the existing tier and any new node capacity required. A free management tier does not mean the workload is free. Consult AKS pricing.

Managed Disk pricing depends on the disk offering. Many offerings use capacity tiers; others expose separately provisioned performance dimensions. Check capacity rounding, IOPS, throughput, transaction or bursting charges where applicable, redundancy choice, and snapshot storage. Consult Managed Disks pricing.

Offer suitable user-provisioned disks where the documented setup supports them. Record subscription, resource group, location, disk ID, SKU, performance settings, and owner. Compare capacity, performance, and access costs before proposing block storage as a midline or farline tier.

Network path

Azure bandwidth pricing distinguishes transfer destinations and regional paths. Check the current availability-zone treatment rather than importing another cloud's cross-zone assumptions. Consult Azure bandwidth pricing.

Inventory the actual route, including peering, NAT, load balancers, private endpoints, or gateways if used. Research each service's own charges; a bandwidth exception does not establish that intermediaries are free. Estimate initial loading, reads, replication, recovery, and tier retrieval in both directions. Keep default storage traffic private and within one region. If the proposed route uses public endpoints or another region, obtain an explicit cost decision before moving evaluation data.

Build the estimate

Separate existing capacity from incremental node runtime, cluster tier charges, managed disks, snapshots, network services, transfer, and monitoring. Record current pricing sources, region, currency, billing units, assumptions, evaluation duration, and cost if left running. Do not assume credits or reserved capacity apply to newly requested resources.

Cleanup evidence

Ask about cleanup at natural stopping points. After authorized teardown, inspect all created resource types, including disks, snapshots, public IPs, network interfaces, load balancers, NAT or gateway resources, and monitoring resources. Check any managed node resource group as well as the evaluation's resource group. Do not delete a shared resource group or pre-existing resources as a shortcut. Report remaining evaluation resources, ongoing costs, and exact cleanup steps.