Cloud SQL Instance vCPU, RAM and Storage
Cloud SQL bills vCPU, memory, and storage as separate SKUs, plus backups. Two things drive surprises: enabling high availability approximately doubles the compute charges, and storage auto-grows but never shrinks.
Billed: Separate per-vCPU-hour, per-gibibyte-hour, and per-gibibyte-month SKUs. High availability roughly doubles the instance charges.
What triggers this charge
- Every hour an instance exists, including instances that are stopped — storage and, for some configurations, other charges continue.
- High availability, which provisions a standby and bills it.
- Automatic storage increases, which are permanent — the disk cannot be shrunk without recreating the instance.
- Backups and point-in-time recovery logs, billed separately from instance storage.
How to reduce it
Right-size, then check HA is actually required
HA on a development or reporting replica doubles the bill for an availability guarantee nobody asked for.
Watch automatic storage growth
It is one-way. A runaway table or an unrotated log can permanently raise the floor of your storage bill.
Buy committed use discounts for steady instances
Cloud SQL supports commitments, and databases are the most predictable workload most organisations run.
Rates
Now find out how much of this you are actually paying.
Finitizer breaks your Google Cloud bill down to the SKU, shows how each one has moved since last week, and separates rate changes from usage changes — so you know whether to talk to engineering or to procurement.