Build capacity reservation configurations for guaranteed compute resources across shapes.
Last verified: May 2026
Build capacity reservation configurations for guaranteed compute resources across shapes.
Required Fields
compartmentIddisplayNameavailabilityDomaininstanceReservationConfigsOutput will appear here...OCI Capacity Reservations guarantee that Compute capacity is available when you need it by reserving specific shapes in specific availability domains. This is critical for workloads that must scale on demand or during planned events where capacity competition could prevent instance launches. This builder helps you configure capacity reservations with shape selection, fault domain preferences, instance counts, and reservation expiry settings.
Capacity reservations in OCI are not charged separately — you only pay for instances that are actually running. However, reserved capacity is guaranteed to be available when you launch instances that match the reservation. This differs from AWS and Azure where reserved capacity may have upfront or ongoing commitment costs. OCI capacity reservations are free to create and maintain; billing only occurs when you consume the reserved capacity by running instances.
Unused reserved capacity remains available for your use at no cost. Other tenants cannot consume your reserved capacity. However, if you consistently reserve capacity you do not use, Oracle may reclaim unused reservations after extended periods of non-use in certain scenarios. It is best practice to match reservations to your actual or projected peak usage and adjust them as requirements change.
Your team is launching a new product expected to generate 5x typical traffic at the launch date. Without capacity reservations, instance launches during the spike could fail if other tenants are also scaling. The builder generates capacity reservations for 50 additional VM.Standard.E5.Flex (8 OCPU, 32 GB) instances spread across all 3 fault domains. Cost: $0/month until launch day. On launch day, instances launch immediately into reserved capacity — no 'sorry, no capacity' errors during the most critical traffic period. After the launch settles (2 weeks later), the reservation is reduced or deleted.
The builder constructs OCI Compute Capacity Reservation configurations: reservation resource (compartment, AD, fault domain optional, instance reservation configs detailing shape and instance count, resource_type=compute-capacity-reservation, instanceReservationConfigs with the shape and count for each fault domain). Output is generated as oci compute compute-capacity-reservation commands and Terraform oci_core_compute_capacity_reservation resources.
OCI Capacity Reservations are FREE — unique advantage vs AWS Reserved Instances or Azure Reserved VMs. You only pay for actually-running instances. Use this aggressively for production workloads, planned events (product launches), and DR site capacity. There's no downside to over-reserving.
Reserve capacity in MULTIPLE fault domains for HA. If you reserve 10 instances all in fault domain 1 and that fault domain has a hardware event, your reserved capacity is unavailable. Spread across all 3 fault domains: 4+3+3 reservation pattern is typical.
GPU shapes are the most capacity-constrained — H100, A100, and other top-tier GPUs have intermittent availability across all clouds. Always reserve GPU capacity in advance. The 'hope it's available' approach for GPU workloads has caused many planned ML training runs to fail.
Was this tool helpful?
Disclaimer: This tool runs entirely in your browser. No data is sent to our servers. Always verify outputs before using them in production. AWS, Azure, and GCP are trademarks of their respective owners.