CostLensDisk Type Downgrade (GCP)
Disk Type Downgrade (GCP)
CostLens flags attached pd-ssd disks with sustained low IOPS that could move to pd-balanced for the same capacity at a lower rate.
Updated July 20261 min read
CostLens flags attached pd-ssd disks with sustained low IOPS. pd-ssd is priced for high IOPS — a disk doing little I/O is paying for performance it doesn't use. Moving it to pd-balanced keeps the same capacity at a lower per-GB rate.
How it works
Disks are enumerated
CostLens lists disks across zones via the Compute aggregated API and keeps the attached pd-ssd ones.
IOPS are measured
Per-device read + write ops are read from Cloud Monitoring; disks below a low ops/sec threshold are flagged.
Savings are computed
Priced as disk size × (pd-ssd rate − pd-balanced rate).
What to do
- Confirm the disk doesn't need pd-ssd's latency for occasional bursts (databases, latency-sensitive apps).
- Change the disk type to pd-balanced (the disk must be detached or the VM stopped to change type).
Severity levels
| Severity | Estimated monthly savings |
|---|---|
| critical | >$500 |
| high | $100–$500 |
| medium | $20–$100 |
| low | <$20 |
Required permissions
- Detect: Compute read + Monitoring Viewer — see GCP Accounts. Without Monitoring read this check produces nothing and never blocks analysis.
Limitations
- IOPS alone can miss latency-sensitivity — review before converting.
- Disks with no Monitoring data are skipped rather than guessed.