Budget AI search and tool usage separately
Separate model tokens, search grounding, and tool-resource measurements so an AI workflow budget does not hide charges behind prompt counts or monitoring data.
Steps3 min read
Cloud cost guides
Find answers to common questions about cloud and AI costs.
58 guides
Review Azure inter-region data transfer costs
Separate bandwidth, peering and Front Door charges before estimating Azure cross-region costs. Use a read-only register to match network paths to billed meters.
Steps3 min read
Check GPU commitment cost against useful work
Separate reservation use from GPU activity, then compare period cost with accepted workload output before deciding whether to renew capacity.
Steps4 min read
Keep shared Azure cost allocation separate from billing
Review shared-service allocation rules, target percentages and cost records so internal chargeback stays separate from Azure invoice reconciliation.
Steps4 min read
Find costly Azure Data Factory activity patterns
Compare factory-level Azure Data Factory costs with per-run meter consumption to choose a workload for review without treating duration or usage as the charged price.
Steps3 min read
Compare AI prices without mixing token types or capacity
Use a read-only worksheet to compare Azure OpenAI, Amazon Bedrock, and Vertex AI costs, separating token categories, cache charges, and provisioned capacity.
Steps4 min read
Avoid double-counting cloud network charges
Separate repeated imports from valid charges on one traffic path. Reconcile a closed invoice period while keeping credits, corrections, adjustments and distinct service fees visible.
Steps4 min read
Check Azure Backup charges before changing retention
Review backup charge types, stopped protection and Archive eligibility before proposing retention changes that could remove needed recovery points.
Steps4 min read
Avoid buying cloud commitments from a single busy week
Before approving a cloud commitment, compare historical recommendations with lasting demand and check whether recent purchases have been reflected.
Steps4 min read
Check discount sharing before comparing project costs
Review Compute Engine discount sharing and fee allocation before treating a lower project cost as lower usage. Include commitment costs outside project totals.
Steps3 min read
Choose an EC2 commitment that fits your workload
Compare Savings Plans and EC2 Reserved Instances with recurring usage, workload changes, and Availability Zone capacity needs before approval.
Steps4 min read
Separate CloudFront charges before tuning the cache
Review CloudFront request and transfer meters, pricing-plan coverage and cache trends to decide whether billing or cache settings need attention.
Steps3 min read
Review Transit Gateway and inter-AZ transfer costs
Check who pays for Transit Gateway attachments and separate hourly, processing and transfer charges before attributing network costs.
BASH4 min read
Check transfer charges before moving Google Cloud workloads
Review Google Cloud network charges after credits, confirm billing units, and link the largest items to traffic a proposed regional move would actually change.
SQL3 min read
Separate Cloud NAT charges from network transfer costs
Break down Cloud NAT costs by gateway hours, external IP hours, processed GiB, and related services. Use existing logs to explain traffic, not to reconstruct the bill.
Steps3 min read
Reduce Athena scans without mistaking alerts for limits
Review Athena partition filters, scan statistics, per-query cancellation and workgroup alerts. Reuse results only when the permitted result age fits the business need.
SQL3 min read
Choose DynamoDB capacity without paying for idle throughput
Compare DynamoDB capacity modes using traffic patterns, CloudWatch metrics, index usage and throttling before changing how you pay for throughput.
Steps4 min read
Separate job trouble from Dataflow cost
Compare Dataflow job estimates with billed costs, then use Dataflow and Managed Service for Apache Spark metrics to assign performance follow-up without mistaking errors for spend.
Steps3 min read
Reduce BigQuery scans without changing the answer
Review one BigQuery SELECT for unnecessary partition scans. Compare dry-run bytes without dropping required rows or treating clustered-table estimates as final costs.
Steps3 min read
Check your BigQuery target before enabling billing export
Review the dataset and project requirements for Cloud Billing export, including location limits, encryption constraints, project links, and available usage-cost history.
Steps3 min read
Review Azure SQL capacity before downsizing
Identify provisioned Azure SQL databases and pools worth a sizing review using workload peaks, storage limits and purchasing-model checks, not low CPU alone.
Steps3 min read
Separate AWS Savings Plans utilization from coverage
Compare consumed commitment with eligible usage coverage in AWS Cost Explorer before deciding whether to investigate unused Savings Plans or review a new purchase.
BASH4 min read
Explain a cloud bill increase before changing resources
Use a read-only worksheet to compare monthly cloud costs and separate usage changes from prices, discounts, credits, and invoice adjustments.
Steps4 min read
Check Lambda memory before cutting it
Use Lambda REPORT logs to spot unused memory, then compare runtime and errors before testing a smaller allocation. Memory headroom alone is not a savings estimate.
Steps3 min read