CloudFinOpsKit

Findings → fixed

The cloud cost remediation playbook

Finding the waste is the easy half. This is how you fix it — every category of Azure and AWS cost finding, mapped to a clear action and the official docs, sequenced so the fastest payback comes first. It's the same plan the CloudFinOpsKit report builds for you, priced and prioritised from your real spend.

Get your remediation plan The FinOps model ↓
  1. 1

    Clear the waste

    Do today · near-zero risk

    Delete orphaned managed disks / EBS volumes, unassociated public IPs, and stale snapshots; stop or deallocate idle VMs, gateways, NAT gateways and load balancers. Nothing in production depends on them — it's immediate, pure saving.

    How to Azure — optimize component costs · AWS — Cost Optimization pillar

  2. 2

    Right-size what's left

    This week · low risk

    Downsize over-provisioned VMs and databases, modernise old SKU/instance generations (incl. Graviton on AWS), enable autoscale, and move dev/test to auto-shutdown or serverless — guided by Azure Advisor and AWS Compute Optimizer.

    How to Azure Advisor cost recommendations · AWS Compute Optimizer

  3. 3

    Commit the steady baseline

    Plan · finance sign-off

    Cover always-on workloads with Reserved Instances or Savings Plans — often 30–70% off pay-as-you-go — and, on Azure, apply Hybrid Benefit to reuse existing Windows/SQL licences. Start with the highest-confidence, steadiest workloads.

    How to Azure reservations · Azure Hybrid Benefit · AWS Savings Plans

  4. 4

    Tame AI & data spend

    This week · low risk

    Delete zombie AI deployments, enable prompt caching, cap output tokens and route simple prompts to cheaper models; right-size provisioned throughput (Azure OpenAI PTU / Bedrock), Log Analytics tiers, log-group retention and backup redundancy (GRS→LRS).

    How to Azure OpenAI provisioned throughput · Amazon Bedrock throughput

  5. 5

    Govern so it can't creep back

    Ongoing · prevents recurrence

    Set budgets with 80/100/120% alerts, enforce required tags with Azure Policy or AWS SCPs, add cost guardrails (allowed SKUs/regions), and re-run monthly to hold the line and watch the FinOps maturity score climb.

    How to Azure budgets & alerts · Azure Policy · AWS Budgets

Run a read-only assessment and get this plan populated with your findings — priced, prioritised, with the exact commands. Browse the templates →

Remediation — FAQ

What should I remediate first?

Clear the waste — orphaned disks/volumes, unassociated IPs, stale snapshots and idle resources. Nothing in production depends on them, so it's immediate saving at near-zero risk.

How do I remediate a rightsizing finding?

Act on Azure Advisor or AWS Compute Optimizer: downsize the over-provisioned VM or database, modernise the SKU/instance family, and schedule non-production off-hours or move it to a serverless tier.

How do I reduce AI and token spend?

Delete unused (zombie) model deployments, enable prompt caching, cap output tokens, route simple prompts to cheaper models, and right-size provisioned throughput (Azure OpenAI PTU / Amazon Bedrock).

How do I stop the waste coming back?

Govern it: budgets with 80/100/120% alerts, required tags enforced by Azure Policy or AWS SCPs, cost guardrails, and a monthly re-run to hold the line.