Google Compute Engine
Virtual machines on Google Cloud, for workloads that need a full operating system rather than a managed platform.
Compute
Chosen per project
Maintainable handover
Compute Engine is the base layer of Google Cloud: a machine with an OS and a network, and nothing decided for you. It is the right home for software that will not fit a managed runtime, and an unnecessary amount of responsibility for software that would.
Two features change the economics compared with a fixed server: per-second billing with sustained-use discounts applied automatically, and custom machine types, so an instance can be sized to the workload instead of rounding up to the next standard shape.
What we build with Compute Engine
Applications with operating-system dependencies
Migrating an existing server with minimal change
Batch and compute-heavy jobs, including preemptible instances
Self-managed databases or middleware
Where it fits — and where it does not
Good fit when
Workloads that need OS control
Steady load, where committed use discounts apply
Teams with someone owning patching
Consider something else when
Standard web applications — Cloud Run is far less work
Spiky workloads paying for idle capacity
Four ways VM spend leaks
Almost every cloud bill we review contains at least two of these.
Oversized shapes
Sized for a guess, never revisited.
Idle instances
Dev machines running through nights and weekends.
Orphaned disks
Persistent disks left behind by deleted VMs.
No commitments
Steady workloads paying on-demand rates.
How we work with Compute Engine
Custom machine types
CPU and memory set to what the workload uses rather than a standard shape that wastes one or the other.
No public IPs by default
Instances in private subnets; access through IAP rather than an open SSH port.
Snapshot schedules
Automated disk snapshots with retention, and a restore proved during the build.
Images, not pets
Machine images built from scripts so an instance can be replaced rather than repaired.
Our typical Compute Engine setup
| Concern | What we use |
|---|---|
| Machine types | Standard or custom shapes, sized from observed usage |
| Networking | VPC with private subnets, Cloud NAT for egress |
| Access | Identity-Aware Proxy for SSH; no public admin ports |
| Backups | Scheduled snapshots with retention and a tested restore |
| Cost | Sustained-use discounts, plus committed use for steady load |
Frequently asked questions
Very closely in capability. The practical differences are custom machine shapes, automatic sustained-use discounts, and which other services you already depend on — that last point usually decides it.
Yes, with licensing included in the instance cost or brought from an existing agreement.
For batch work that can be interrupted and resumed, they are dramatically cheaper. For anything user-facing, no.
Topics people search for
Services built with Compute Engine
Cloud Solutions
Hosting designed around what your application actually needs, with the monitoring and backups that make it safe to rely on.
Custom Software Development
Software shaped around the way your business already works, instead of a packaged product you have to reorganise around.
Software Maintenance
Keeping working systems working — including ones built by somebody else.
Enterprise Software Development
Systems for organisations where several departments, sites and approval chains all have to work from the same data.
Tell us what you are trying to build
Describe the problem in plain language and we will tell you what it would take to solve it — the approach, the moving parts and the sensible order to build them in. No obligation either way.