Optimising Cloud Costs | FinOps for GCP Migration

A “Lift-and-Shift” migration often leads to a nasty surprise when the first bill arrives. This is because on-premise servers are usually bigger than they need to be to handle occasional spikes in traffic. Moving these oversized servers straight to the cloud means you are paying for capacity you aren’t using. To avoid “Bill Shock,” you need to think about FinOps from Day 1.

1. Rightsizing: The Low Hanging Fruit

Before you move a single server, use a tool like Google StratoZone to check what you are actually using.

Scenario: You have a massive 64-core server, but it’s only using 10% of its power on average.
Action: Move it to a much smaller 8-core or 16-core server in the cloud. This simple step alone can save you 50-70% on your bill.

2. Using Cloud Pricing to Your Advantage

The cloud offers savings that hardware vendors just can’t match:

Spot VMs

For tasks that can be interrupted (like processing a batch of images), Google offers spare capacity at up to 80% discount. We build reliable systems on top of these cheap, unreliable nodes to get the best of both worlds.

Committed Use Discounts (CUDs)

If you know you’re going to use a certain amount of computer power for the next year, you can commit to it in exchange for a big discount. Unlike some other clouds, Google allows you to apply this discount flexibly across different machine types.

3. Smarter Storage

Data Lakes often contain Petabytes of data that nobody has looked at in years. We set up automatic rules on Google Cloud Storage buckets:

Day 0: Keep data in “Standard” storage for fast access.
Day 30: Move it to “Nearline” storage to save money.
Day 90: Move it to “Archive” storage for long-term keeping at the lowest price.

4. Tagging and Accountability

You can’t save money if you don’t know who is spending it. We implement strict labeling rules (e.g., tagging every server with a Project and CostCenter). This lets the Finance team see exactly where the money is going.

Alps Agility guarantees that our architectures are cost-optimised by design. We help you shift from expensive capital expenditure to a highly efficient operating model.