Rename ci-training → ci-training-l40s, add H100/H100x2/H100-SXM/H100-SXM-8 pool resources (all autoscaling 0→1). Remove dev pool (DevPod on platform). Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
168 lines
4.3 KiB
HCL
168 lines
4.3 KiB
HCL
variable "region" {
|
||
description = "Scaleway region for the Kapsule cluster"
|
||
type = string
|
||
}
|
||
|
||
variable "cluster_name" {
|
||
description = "Name of the Kubernetes cluster"
|
||
type = string
|
||
default = "foxhunt"
|
||
}
|
||
|
||
variable "k8s_version" {
|
||
description = "Kubernetes version for the Kapsule cluster"
|
||
type = string
|
||
default = "1.34"
|
||
}
|
||
|
||
variable "foxhunt_type" {
|
||
description = "Instance type for the foxhunt app services node pool"
|
||
type = string
|
||
default = "DEV1-L"
|
||
}
|
||
|
||
variable "foxhunt_max_size" {
|
||
description = "Maximum number of nodes in the foxhunt pool (autoscaling)"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_platform_pool" {
|
||
description = "Create dedicated node pool for platform infrastructure (databases, monitoring)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "platform_type" {
|
||
description = "Instance type for the platform infrastructure node pool"
|
||
type = string
|
||
default = "DEV1-L"
|
||
}
|
||
|
||
variable "enable_gitlab_pool" {
|
||
description = "Create dedicated node pool for GitLab CE"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "gitlab_type" {
|
||
description = "Instance type for the GitLab node pool"
|
||
type = string
|
||
default = "DEV1-L"
|
||
}
|
||
|
||
variable "enable_ci_training_l40s_pool" {
|
||
description = "Create L40S GPU node pool for training + hyperopt (48GB VRAM, CC 89)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_training_l40s_type" {
|
||
description = "Instance type for the L40S training pool"
|
||
type = string
|
||
default = "L40S-1-48G"
|
||
}
|
||
|
||
variable "ci_training_l40s_max_size" {
|
||
description = "Maximum number of nodes in the L40S training pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_ci_training_h100_pool" {
|
||
description = "Create H100 GPU node pool for training + hyperopt (80GB VRAM, CC 90)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_training_h100_type" {
|
||
description = "Instance type for the H100 training pool"
|
||
type = string
|
||
default = "H100-1-80G"
|
||
}
|
||
|
||
variable "ci_training_h100_max_size" {
|
||
description = "Maximum number of nodes in the H100 training pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_ci_training_h100x2_pool" {
|
||
description = "Create H100x2 GPU node pool for multi-GPU training (2× 80GB VRAM)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_training_h100x2_type" {
|
||
description = "Instance type for the H100x2 training pool"
|
||
type = string
|
||
default = "H100-2-80G"
|
||
}
|
||
|
||
variable "ci_training_h100x2_max_size" {
|
||
description = "Maximum number of nodes in the H100x2 training pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_ci_training_h100_sxm_pool" {
|
||
description = "Create H100-SXM GPU node pool for NVLink multi-GPU training (2× 80GB SXM)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_training_h100_sxm_type" {
|
||
description = "Instance type for the H100-SXM training pool"
|
||
type = string
|
||
default = "H100-SXM-2-80G"
|
||
}
|
||
|
||
variable "ci_training_h100_sxm_max_size" {
|
||
description = "Maximum number of nodes in the H100-SXM training pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_ci_training_h100_sxm8_pool" {
|
||
description = "Create H100-SXM-8 GPU node pool for full multi-GPU training (8× 80GB SXM, NVLink)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_training_h100_sxm8_type" {
|
||
description = "Instance type for the H100-SXM-8 training pool"
|
||
type = string
|
||
default = "H100-SXM-8-80G"
|
||
}
|
||
|
||
variable "ci_training_h100_sxm8_max_size" {
|
||
description = "Maximum number of nodes in the H100-SXM-8 training pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "enable_ci_compile_cpu_pool" {
|
||
description = "Create high-CPU node pool for CI compilation (no GPU needed)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|
||
variable "ci_compile_cpu_type" {
|
||
description = "Instance type for the CPU compile pool (32 vCPU, 128GB RAM, no GPU)"
|
||
type = string
|
||
default = "POP2-32C-128G"
|
||
}
|
||
|
||
variable "ci_compile_cpu_max_size" {
|
||
description = "Maximum number of nodes in the CPU compile pool"
|
||
type = number
|
||
default = 1
|
||
}
|
||
|
||
variable "public_ip_disabled" {
|
||
description = "Disable public IPs on all node pools (requires Public Gateway for egress)"
|
||
type = bool
|
||
default = false
|
||
}
|
||
|