Manage Storage IOPS usage
What you are charged for
On the Platform API, IOPS (Input/Output Operations Per Second) represents the provisioned random read/write performance capacity of AWS-backed block storage used by any workload.
This includes, but is not limited to:
- Database engines (PostgreSQL and other SQL systems)
- Stateful compute workloads with persistent storage
- Background job systems using attached volumes
- Cache or indexing layers using block storage
- Any AWS EBS-backed storage volume
IOPS is provisioned per workload and mapped directly to AWS storage performance configuration.
Infrastructure models:
- Silo – dedicated AWS infrastructure per tenant (isolated compute + storage performance)
- Bridge – shared infrastructure with logically isolated IOPS allocation per tenant
- Pool – fully shared multi-tenant infrastructure with pooled IOPS capacity
How charges are calculated
IOPS is billed in IOPS-Hours (IOPS-Hrs):
1 IOPS-Hr = 1 provisioned IOPS for 1 hour
Example:
- 1,000 IOPS for 5 hours = 5,000 IOPS-Hrs
- 3,600 IOPS for 10 hours = 36,000 IOPS-Hrs
Charges apply only to provisioned IOPS above the included baseline where applicable.
Usage on your invoice
IOPS usage is shown as:
- Storage IOPS-Hrs
Pricing model
IOPS pricing is derived from AWS EBS performance tiers (gp3 baseline + io2 provisioned IOPS) plus Platform API orchestration overhead.
Pricing varies by disk type. The default is gp3 (AWS General Purpose SSD).
General purpose storage (gp3)
AWS gp3 includes a baseline IOPS allocation:
- $0.00003288 per IOPS-Hr (~$0.024 per IOPS per month)
- Includes baseline of 3,000 IOPS per workload
You are only charged for provisioned IOPS exceeding this baseline.
| Plan | Included IOPS | Over-Usage Pricing Model |
|---|---|---|
| Silo | 3,000 | AWS gp3 rate |
| Bridge | 3,000 | AWS gp3 rate |
| Pool | 3,000 (logical) | AWS gp3 rate |
| Enterprise | Custom | Custom AWS agreement |
High performance storage (io2)
AWS io2 provides fully provisioned IOPS:
- $0.000163 per IOPS-Hr (~$0.119 per IOPS per month)
- No included baseline (billed from first provisioned IOPS)
| Plan | Included IOPS | Pricing Model |
|---|---|---|
| Silo | 0 | AWS io2 rate |
| Bridge | 0 | AWS io2 rate |
| Pool | 0 | AWS io2 rate |
| Enterprise | Custom | Custom AWS agreement |
Tenancy impact on IOPS
IOPS behavior depends on infrastructure model:
Silo (Dedicated)
- Dedicated AWS storage performance allocation
- Predictable latency and throughput under load
- Strong isolation of noisy neighbors
Bridge (Hybrid)
- Shared infrastructure with isolated IOPS reservations per tenant
- Balanced cost vs performance
Pool (Shared)
- Fully shared IOPS pool across workloads
- Dynamic allocation based on demand and fairness policies
Billing examples
gp3 IOPS
Only IOPS above the included baseline (3,000 IOPS) is billed.
| Line Item | Units | Cost |
|---|---|---|
| Compute (Silo) | 730 hours | $25 |
| Compute (Bridge) | 730 hours | $25 |
| Compute (Pool) | 730 hours | $25 |
| Storage IOPS Project 1 | 3,000 IOPS | $0 |
| Storage IOPS Project 2 | 3,600 IOPS | $14.40 |
| Subtotal | $165 | |
| Credits | -$10 | |
| Total | $155 |
io2 IOPS
io2 is billed from the first provisioned IOPS.
| Line Item | Units | Cost |
|---|---|---|
| Compute (Large Workload) | 730 hours | $111 |
| Storage IOPS Workload | 8,000 IOPS | $952 |
| Subtotal | $1,088 | |
| Credits | -$10 | |
| Total | $1,078 |
Key behavior notes
- IOPS applies to any stateful AWS-backed workload, not only databases
- Each workload instance has its own IOPS profile
- Silo = fixed dedicated performance
- Bridge = reserved shared performance
- Pool = dynamic shared performance allocation
View usage
You can monitor IOPS usage in the Platform API control plane:
Organization → Usage → Storage Performance
The dashboard shows:
- Per-project IOPS allocation
- Breakdown by Silo / Bridge / Pool
- Real-time performance consumption
- Historical IOPS utilization trends
Optimize usage
- Avoid over-provisioning IOPS for low-traffic workloads
- Use Pool for bursty or unpredictable workloads
- Reserve Silo IOPS only for latency-critical systems
- Tune database indexes and query patterns to reduce random I/O
- Batch writes to reduce sustained IOPS demand
- Consolidate low-utilization workloads under shared infrastructure
Exceeding quotas
When IOPS usage exceeds included allocations:
- Silo / Bridge / Pool (paid without spend caps): billed automatically at AWS io2/gp3 rates
- Paid plans with spend caps enabled: IOPS provisioning may be throttled or limited
All billing is directly derived from AWS EBS IOPS pricing and region-specific capacity constraints.