When a business runs workloads on more than one public cloud, spending becomes harder to see and harder to control. This guide outlines a simple governance model that works across Azure, Alibaba Cloud and Tencent Cloud.

Running workloads on more than one public cloud gives teams flexibility: they can place each application where the services, regions or commercial terms suit it. The trade-off is visibility. Each provider has its own billing console, its own pricing units and its own discount programmes, so a simple question such as "what did we spend on our e-commerce platform last month?" can take days to answer.

Cost governance does not need a large team or an expensive tool to get started. It needs clear ownership, consistent data and a regular review rhythm. The model below follows the three phases described by the FinOps community: Inform, Optimize and Operate.

1. Inform: make spending visible

The first goal is to answer "who spends what, and why" for every cloud account you hold.

  • Agree a tagging standard. Choose a short list of mandatory tags (or labels) such as owner, application, environment and cost-centre, and use the same names on every platform.
  • Map accounts to the business. Keep a register that links each Azure subscription, Alibaba Cloud account and Tencent Cloud account to a business unit and a named owner.
  • Normalise the data. Export billing data from each provider on a schedule and convert it into one format with common fields: date, account, service, region, tags, currency and amount.
  • Separate shared costs. Network egress, support plans and shared security tooling rarely carry application tags. Decide an allocation rule up front and apply it consistently.

2. Optimize: remove waste before negotiating discounts

It is tempting to start with commitment discounts, but committing to capacity you do not need simply locks in waste. A sensible order of work is:

  1. Clean up idle resources. Unattached disks, unused public IP addresses, stopped virtual machines that still reserve compute, and empty application service plans are common findings on every platform.
  2. Right-size. Compare provisioned CPU and memory with observed utilisation over at least two to four weeks before changing instance sizes.
  3. Schedule non-production workloads. Development and test environments often do not need to run overnight or at weekends.
  4. Then consider commitments. Reserved capacity and savings plans can reduce unit prices for steady workloads. Terms, scope and exchange rules differ between providers, so review them carefully before purchasing.
Tip: record the expected monthly saving and the date of every optimisation action. Comparing complete billing periods before and after the change is a reliable way to confirm the saving.

3. Operate: build a monthly rhythm

  • Set budgets and alerts for each account and each major application.
  • Hold a short monthly review with application owners and finance. Look at the top cost changes, open optimisation actions and upcoming commitment renewals.
  • Add cost checks to change management: new high-cost resources should be visible to the owner on the day they are created, not at the end of the month.

Where a cloud reseller can help

A reseller that manages accounts on several platforms can consolidate invoices, help apply a consistent tagging and account structure, and advise on the commercial programmes each provider offers. The governance decisions themselves, such as who owns each application and how shared costs are allocated, remain with your business.

This article is general information and not legal, tax or financial advice. Service features, regions and prices change; confirm current details with each provider before making decisions.