Skip to main content
resource · gcp

Compute Engine Instance

live rule families
13
schedulable
yes
category
compute-services

Does ZopNight manage Compute Engine Instance?

Compute Engine bills vCPUs and memory per second with a 1-minute minimum, and stopping an instance ends those meters while persistent disks and static IPs keep billing. ZopNight discovers every VM via Cloud Asset Inventory, watches CPU and memory over a 42-day lookback, and stops instances on schedule.

At a glance

Compute Engine Instance coverage facts.
Field Value
Scheduling notesissues an instance stop, releasing vCPU and memory billing while preserving disks; restarts on schedule

Compute Engine virtual machines are the workhorse of GCP, billed per second for vCPUs, memory, GPUs, and attached resources. VMs that run 24x7 for workloads needed only during business hours typically waste 65 to 75 percent of their compute spend.

Per-second meters and the discounts layered on them

Compute Engine charges separate per-second meters (with a 1-minute minimum) for vCPUs, memory, and any attached GPUs, at rates set by machine family and region. Two discount mechanisms behave very differently under scheduling. Sustained-use discounts apply automatically as an instance runs more of the month, so a stopped VM simply earns less discount on a much smaller bill, which is still a clear win. Committed use discounts bill for the committed capacity regardless of whether anything runs, so stopping a VM covered by a commitment saves nothing until the commitment is re-used elsewhere.

How ZopNight covers the fleet

ZopNight discovers every VM via Cloud Asset Inventory, monitors CPU and memory over a 42-day lookback, rightsizes through 128 GCP recommendation rules, and stops instances on schedule. Managed-group members are handled through the MIG rather than individually, because an autoscaler would immediately recreate an instance stopped behind its back.

What a stop ends, and what it does not

Stopping an instance releases its vCPU and memory billing while preserving its persistent disks. The disks keep billing by provisioned size the entire time the VM is stopped, as does any reserved static external IP. On a small machine with a large disk, the stopped-state cost can be a substantial fraction of the running cost. Local SSDs are the sharp edge: their data does not survive a stop.

Where Compute Engine money leaks

The recurring patterns: development and staging VMs running all 168 hours of the week for a 45-hour working window; machines sized for a launch-day peak that never returned; and GPUs attached to instances whose workload finished weeks ago, still metering per second.

Console path

Google Cloud console → Compute Engine → VM instances lists every instance with its status. A stopped instance shows Terminated: alarming wording, but it means exactly the billing-off state a schedule aims for.

See it fire on your bill.

Connect an account read-only. The first findings land in minutes.

417 rule families across 353 resource types on 22 platforms. Every threshold, metric, and IAM action is documented on these pages before you grant anything.

417 rule families documented
353 resource types covered
read-only default access level
Multi-cloud automation· Production-ready in 30 min· SOC 2 · ISO 27001· 20–60% off the bill, first month· 4 platforms · 1 console· Multi-cloud automation· Production-ready in 30 min· SOC 2 · ISO 27001· 20–60% off the bill, first month· 4 platforms · 1 console·