{ "title": "DstackConfigurationRequest", "discriminator": { "propertyName": "type", "mapping": { "dev-environment": "#/definitions/DevEnvironmentConfigurationRequest", "task": "#/definitions/TaskConfigurationRequest", "service": "#/definitions/ServiceConfigurationRequest", "fleet": "#/definitions/FleetConfigurationRequest", "gateway": "#/definitions/GatewayConfigurationRequest", "volume": "#/definitions/VolumeConfigurationRequest" } }, "oneOf": [ { "$ref": "#/definitions/DevEnvironmentConfigurationRequest" }, { "$ref": "#/definitions/TaskConfigurationRequest" }, { "$ref": "#/definitions/ServiceConfigurationRequest" }, { "$ref": "#/definitions/FleetConfigurationRequest" }, { "$ref": "#/definitions/GatewayConfigurationRequest" }, { "$ref": "#/definitions/VolumeConfigurationRequest" } ], "additionalProperties": true, "$schema": "http://json-schema.org/draft-07/schema#", "definitions": { "PortMappingRequest": { "title": "PortMappingRequest", "type": "object", "properties": { "local_port": { "title": "Local Port", "exclusiveMinimum": 0, "maximum": 65536, "type": "integer" }, "container_port": { "title": "Container Port", "exclusiveMinimum": 0, "maximum": 65536, "type": "integer" } }, "required": [ "container_port" ], "additionalProperties": false }, "RegistryAuthRequest": { "title": "RegistryAuthRequest", "description": "Credentials for pulling a private Docker image.\n\nAttributes:\n username (str): The username\n password (str): The password or access token", "type": "object", "properties": { "username": { "title": "Username", "description": "The username", "type": "string" }, "password": { "title": "Password", "description": "The password or access token", "type": "string" } }, "required": [ "username", "password" ], "additionalProperties": false }, "PythonVersion": { "title": "PythonVersion", "description": "An enumeration.", "enum": [ "3.9", "3.10", "3.11", "3.12", "3.13" ], "type": "string" }, "EnvSentinelRequest": { "title": "EnvSentinelRequest", "type": "object", "properties": { "key": { "title": "Key", "type": "string" } }, "required": [ "key" ], "additionalProperties": false }, "Env": { "title": "Env", "description": "Env represents a mapping of process environment variables, as in environ(7).\nEnvironment values may be omitted, in that case the :class:`EnvSentinel`\nobject is used as a placeholder.\n\nTo create an instance from a `dict[str, str]` or a `list[str]` use pydantic's\n:meth:`BaseModel.parse_obj(dict | list)` method.\n\nNB: this is *NOT* a CoreModel, pydantic-duality, which is used as a base\nfor the CoreModel, doesn't play well with custom root models.", "default": {}, "anyOf": [ { "type": "array", "items": { "type": "string" } }, { "type": "object", "additionalProperties": { "anyOf": [ { "type": "string" }, { "$ref": "#/definitions/EnvSentinelRequest" } ] } } ] }, "CPUArchitecture": { "title": "CPUArchitecture", "description": "An enumeration.", "enum": [ "x86", "arm" ], "type": "string" }, "Range_int_": { "title": "Range[int]", "type": "object", "properties": { "min": { "title": "Min", "type": "integer" }, "max": { "title": "Max", "type": "integer" } }, "additionalProperties": false }, "CPUSpecRequest": { "title": "CPUSpecRequest", "type": "object", "properties": { "arch": { "description": "The CPU architecture, one of: `x86`, `arm`", "allOf": [ { "$ref": "#/definitions/CPUArchitecture" } ] }, "count": { "title": "Count", "description": "The number of CPU cores", "default": { "min": 2, "max": null }, "anyOf": [ { "$ref": "#/definitions/Range_int_" }, { "type": "integer" }, { "type": "string" } ] } }, "additionalProperties": false }, "Range_Memory_": { "title": "Range[Memory]", "type": "object", "properties": { "min": { "title": "Min", "type": "number" }, "max": { "title": "Max", "type": "number" } }, "additionalProperties": false }, "AcceleratorVendor": { "title": "AcceleratorVendor", "description": "An enumeration.", "enum": [ "nvidia", "amd", "google", "intel", "tenstorrent" ], "type": "string" }, "GPUSpecRequest": { "title": "GPUSpecRequest", "type": "object", "properties": { "vendor": { "description": "The vendor of the GPU/accelerator, one of: `nvidia`, `amd`, `google` (alias: `tpu`), `intel`", "allOf": [ { "$ref": "#/definitions/AcceleratorVendor" } ] }, "name": { "title": "Name", "description": "The name of the GPU (e.g., `A100` or `H100`)", "items": { "type": "string" }, "anyOf": [ { "type": "array" }, { "type": "string" } ] }, "count": { "title": "Count", "description": "The number of GPUs", "default": { "min": 1, "max": null }, "anyOf": [ { "$ref": "#/definitions/Range_int_" }, { "type": "integer" }, { "type": "string" } ] }, "memory": { "title": "Memory", "description": "The RAM size (e.g., `16GB`). Can be set to a range (e.g. `16GB..`, or `16GB..80GB`)", "anyOf": [ { "$ref": "#/definitions/Range_Memory_" }, { "type": "integer" }, { "type": "string" } ] }, "total_memory": { "title": "Total Memory", "description": "The total RAM size (e.g., `32GB`). Can be set to a range (e.g. `16GB..`, or `16GB..80GB`)", "anyOf": [ { "$ref": "#/definitions/Range_Memory_" }, { "type": "integer" }, { "type": "string" } ] }, "compute_capability": { "title": "Compute Capability", "description": "The minimum compute capability of the GPU (e.g., `7.5`)", "type": "array", "items": {} } }, "additionalProperties": false }, "DiskSpecRequest": { "title": "DiskSpecRequest", "type": "object", "properties": { "size": { "title": "Size", "description": "Disk size", "anyOf": [ { "$ref": "#/definitions/Range_Memory_" }, { "type": "integer" }, { "type": "string" } ] } }, "required": [ "size" ], "additionalProperties": false }, "ResourcesSpecRequest": { "title": "ResourcesSpecRequest", "type": "object", "properties": { "cpu": { "title": "Cpu", "description": "The CPU requirements", "default": { "arch": null, "count": { "min": 2, "max": null } }, "anyOf": [ { "$ref": "#/definitions/CPUSpecRequest" }, { "$ref": "#/definitions/Range_int_" }, { "type": "integer" }, { "type": "string" } ] }, "memory": { "title": "Memory", "description": "The RAM size (e.g., `8GB`)", "default": { "min": 8.0, "max": null }, "anyOf": [ { "$ref": "#/definitions/Range_Memory_" }, { "type": "integer" }, { "type": "string" } ] }, "shm_size": { "title": "Shm Size", "description": "The size of shared memory (e.g., `8GB`). If you are using parallel communicating processes (e.g., dataloaders in PyTorch), you may need to configure this", "anyOf": [ { "type": "number" }, { "type": "integer" }, { "type": "string" } ] }, "gpu": { "title": "Gpu", "description": "The GPU requirements", "default": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "anyOf": [ { "$ref": "#/definitions/GPUSpecRequest" }, { "type": "integer" }, { "type": "string" } ] }, "disk": { "title": "Disk", "description": "The disk resources", "default": { "size": { "min": 100.0, "max": null } }, "anyOf": [ { "$ref": "#/definitions/DiskSpecRequest" }, { "type": "integer" }, { "type": "string" } ] } }, "additionalProperties": false }, "VolumeMountPointRequest": { "title": "VolumeMountPointRequest", "type": "object", "properties": { "name": { "title": "Name", "description": "The network volume name or the list of network volume names to mount. If a list is specified, one of the volumes in the list will be mounted. Specify volumes from different backends/regions to increase availability", "anyOf": [ { "type": "string" }, { "type": "array", "items": { "type": "string" } } ] }, "path": { "title": "Path", "description": "The absolute container path to mount the volume at", "type": "string" } }, "required": [ "name", "path" ], "additionalProperties": false }, "InstanceMountPointRequest": { "title": "InstanceMountPointRequest", "type": "object", "properties": { "instance_path": { "title": "Instance Path", "description": "The absolute path on the instance (host)", "type": "string" }, "path": { "title": "Path", "description": "The absolute path in the container", "type": "string" }, "optional": { "title": "Optional", "description": "Allow running without this volume in backends that do not support instance volumes", "default": false, "type": "boolean" } }, "required": [ "instance_path", "path" ], "additionalProperties": false }, "RepoExistsAction": { "title": "RepoExistsAction", "description": "An enumeration.", "enum": [ "error", "skip" ], "type": "string" }, "RepoSpecRequest": { "title": "RepoSpecRequest", "type": "object", "properties": { "local_path": { "title": "Local Path", "description": "The path to the Git repo on the user's machine. Relative paths are resolved relative to the parent directory of the the configuration file. Mutually exclusive with `url`", "type": "string" }, "url": { "title": "Url", "description": "The Git repo URL. Mutually exclusive with `local_path`", "type": "string" }, "branch": { "title": "Branch", "description": "The repo branch. Defaults to the active branch for local paths and the default branch for URLs", "type": "string" }, "hash": { "title": "Hash", "description": "The commit hash", "type": "string" }, "path": { "title": "Path", "description": "The repo path inside the run container. Relative paths are resolved relative to the working directory", "default": ".", "type": "string" }, "if_exists": { "description": "The action to be taken if `path` exists and is not empty. One of: `error`, `skip`", "default": "error", "allOf": [ { "$ref": "#/definitions/RepoExistsAction" } ] } }, "additionalProperties": false }, "FilePathMappingRequest": { "title": "FilePathMappingRequest", "type": "object", "properties": { "local_path": { "title": "Local Path", "description": "The path on the user's machine. Relative paths are resolved relative to the parent directory of the the configuration file", "type": "string" }, "path": { "title": "Path", "description": "The path in the container. Relative paths are resolved relative to the working directory", "type": "string" } }, "required": [ "local_path", "path" ], "additionalProperties": false }, "BackendType": { "title": "BackendType", "description": "Attributes:\n AMDDEVCLOUD (BackendType): AMD Developer Cloud\n AWS (BackendType): Amazon Web Services\n AZURE (BackendType): Microsoft Azure\n CLOUDRIFT (BackendType): CloudRift\n CRUSOE (BackendType): Crusoe\n CUDO (BackendType): Cudo\n DATACRUNCH (BackendType): DataCrunch (for backward compatibility)\n DIGITALOCEAN (BackendType): DigitalOcean\n DSTACK (BackendType): dstack Sky\n GCP (BackendType): Google Cloud Platform\n HOTAISLE (BackendType): Hot Aisle\n KUBERNETES (BackendType): Kubernetes\n LAMBDA (BackendType): Lambda Cloud\n NEBIUS (BackendType): Nebius AI Cloud\n OCI (BackendType): Oracle Cloud Infrastructure\n RUNPOD (BackendType): Runpod Cloud\n TENSORDOCK (BackendType): TensorDock Marketplace\n VASTAI (BackendType): Vast.ai Marketplace\n VERDA (BackendType): Verda Cloud\n VULTR (BackendType): Vultr", "enum": [ "amddevcloud", "aws", "azure", "cloudrift", "crusoe", "cudo", "datacrunch", "digitalocean", "dstack", "gcp", "hotaisle", "kubernetes", "lambda", "local", "remote", "nebius", "oci", "runpod", "tensordock", "vastai", "verda", "vultr" ], "type": "string" }, "SpotPolicy": { "title": "SpotPolicy", "description": "An enumeration.", "enum": [ "spot", "on-demand", "auto" ], "type": "string" }, "RetryEvent": { "title": "RetryEvent", "description": "An enumeration.", "enum": [ "no-capacity", "interruption", "error" ], "type": "string" }, "ProfileRetryRequest": { "title": "ProfileRetryRequest", "type": "object", "properties": { "on_events": { "description": "The list of events that should be handled with retry. Supported events are `no-capacity`, `interruption`, `error`. Omit to retry on all events", "type": "array", "items": { "$ref": "#/definitions/RetryEvent" } }, "duration": { "title": "Duration", "description": "The maximum period of retrying the run, e.g., `4h` or `1d`. The period is calculated as a run age for `no-capacity` event and as a time passed since the last `interruption` and `error` for `interruption` and `error` events.", "anyOf": [ { "type": "integer" }, { "type": "string" } ] } }, "additionalProperties": false }, "CreationPolicy": { "title": "CreationPolicy", "description": "An enumeration.", "enum": [ "reuse", "reuse-or-create" ], "type": "string" }, "UtilizationPolicyRequest": { "title": "UtilizationPolicyRequest", "type": "object", "properties": { "min_gpu_utilization": { "title": "Min Gpu Utilization", "description": "Minimum required GPU utilization, percent. If any GPU has utilization below specified value during the whole time window, the run is terminated", "minimum": 0, "maximum": 100, "type": "integer" }, "time_window": { "title": "Time Window", "description": "The time window of metric samples taking into account to measure utilization (e.g., `30m`, `1h`). Minimum is `5m`", "anyOf": [ { "type": "integer" }, { "type": "string" } ] } }, "required": [ "min_gpu_utilization", "time_window" ], "additionalProperties": false }, "StartupOrder": { "title": "StartupOrder", "description": "An enumeration.", "enum": [ "any", "master-first", "workers-first" ], "type": "string" }, "StopCriteria": { "title": "StopCriteria", "description": "An enumeration.", "enum": [ "all-done", "master-done" ], "type": "string" }, "ScheduleRequest": { "title": "ScheduleRequest", "type": "object", "properties": { "cron": { "title": "Cron", "description": "A cron expression or a list of cron expressions specifying the UTC time when the run needs to be started", "anyOf": [ { "type": "array", "items": { "type": "string" } }, { "type": "string" } ] } }, "required": [ "cron" ], "additionalProperties": false }, "EntityReferenceRequest": { "title": "EntityReferenceRequest", "description": "Cross-project entity reference.", "type": "object", "properties": { "project": { "title": "Project", "description": "The project name. If unspecified, refers to the current project", "type": "string" }, "name": { "title": "Name", "description": "The entity name", "type": "string" } }, "required": [ "name" ], "additionalProperties": false }, "DevEnvironmentConfigurationRequest": { "title": "DevEnvironmentConfigurationRequest", "type": "object", "properties": { "ide": { "title": "Ide", "description": "The IDE to pre-install. Supported values include `vscode`, `cursor`, and `windsurf`. Defaults to no IDE (SSH only)", "anyOf": [ { "enum": [ "vscode" ], "type": "string" }, { "enum": [ "cursor" ], "type": "string" }, { "enum": [ "windsurf" ], "type": "string" } ] }, "version": { "title": "Version", "description": "The version of the IDE. For `windsurf`, the version is in the format `version@commit`", "type": "string" }, "init": { "title": "Init", "description": "The shell commands to run on startup", "default": [], "type": "array", "items": { "type": "string" } }, "inactivity_duration": { "title": "Inactivity Duration", "description": "The maximum amount of time the dev environment can be inactive (e.g., `2h`, `1d`, etc). After it elapses, the dev environment is automatically stopped. Inactivity is defined as the absence of SSH connections to the dev environment, including VS Code connections, `ssh ` shells, and attached `dstack apply` or `dstack attach` commands. Use `off` for unlimited duration. Can be updated in-place. Defaults to `off`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "ports": { "title": "Ports", "description": "Port numbers/mapping to expose", "default": [], "type": "array", "items": { "anyOf": [ { "type": "integer", "exclusiveMinimum": 0, "maximum": 65536 }, { "type": "string", "pattern": "^(?:[0-9]+|\\*):[0-9]+$" }, { "$ref": "#/definitions/PortMappingRequest" } ] } }, "type": { "title": "Type", "default": "dev-environment", "enum": [ "dev-environment" ], "type": "string" }, "name": { "title": "Name", "description": "The run name. If not specified, a random name is generated", "type": "string" }, "image": { "title": "Image", "description": "The name of the Docker image to run", "type": "string" }, "user": { "title": "User", "description": "The user inside the container, `user_name_or_id[:group_name_or_id]` (e.g., `ubuntu`, `1000:1000`). Defaults to the default user from the `image`", "type": "string" }, "privileged": { "title": "Privileged", "description": "Run the container in privileged mode", "default": false, "type": "boolean" }, "entrypoint": { "title": "Entrypoint", "description": "The Docker entrypoint", "type": "string" }, "working_dir": { "title": "Working Dir", "description": "The absolute path to the working directory inside the container. Defaults to the `image`'s default working directory", "type": "string" }, "home_dir": { "title": "Home Dir", "default": "/root", "type": "string" }, "registry_auth": { "title": "Registry Auth", "description": "Credentials for pulling a private Docker image", "allOf": [ { "$ref": "#/definitions/RegistryAuthRequest" } ] }, "python": { "description": "The major version of Python. Mutually exclusive with `image` and `docker`", "allOf": [ { "$ref": "#/definitions/PythonVersion" } ] }, "nvcc": { "title": "Nvcc", "description": "Use image with NVIDIA CUDA Compiler (NVCC) included. Mutually exclusive with `image` and `docker`", "type": "boolean" }, "single_branch": { "title": "Single Branch", "description": "Whether to clone and track only the current branch or all remote branches. Relevant only when using remote Git repos. Defaults to `false` for dev environments and to `true` for tasks and services", "type": "boolean" }, "env": { "title": "Env", "description": "The mapping or the list of environment variables", "default": { "__root__": {} }, "allOf": [ { "$ref": "#/definitions/Env" } ] }, "shell": { "title": "Shell", "description": "The shell used to run commands. Allowed values are `sh`, `bash`, or an absolute path, e.g., `/usr/bin/zsh`. Defaults to `/bin/sh` if the `image` is specified, `/bin/bash` otherwise", "type": "string" }, "resources": { "title": "Resources", "description": "The resources requirements to run the configuration", "default": { "cpu": { "min": 2, "max": null }, "memory": { "min": 8.0, "max": null }, "shm_size": null, "gpu": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "disk": { "size": { "min": 100.0, "max": null } } }, "allOf": [ { "$ref": "#/definitions/ResourcesSpecRequest" } ] }, "priority": { "title": "Priority", "description": "The priority of the run, an integer between `0` and `100`. `dstack` tries to provision runs with higher priority first. Defaults to `0`", "minimum": 0, "maximum": 100, "type": "integer" }, "volumes": { "title": "Volumes", "description": "The volumes mount points", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/VolumeMountPointRequest" }, { "$ref": "#/definitions/InstanceMountPointRequest" }, { "type": "string" } ] } }, "docker": { "title": "Docker", "description": "Use Docker inside the container. Mutually exclusive with `image`, `python`, and `nvcc`. Overrides `privileged`", "type": "boolean" }, "repos": { "title": "Repos", "description": "The list of Git repos", "default": [], "type": "array", "items": { "$ref": "#/definitions/RepoSpecRequest" } }, "files": { "title": "Files", "description": "The local to container file path mappings", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/FilePathMappingRequest" }, { "type": "string" } ] } }, "setup": { "title": "Setup", "default": [], "type": "array", "items": { "type": "string" } }, "backends": { "description": "The backends to consider for provisioning (e.g., `[aws, gcp]`)", "type": "array", "items": { "$ref": "#/definitions/BackendType" } }, "regions": { "title": "Regions", "description": "The regions to consider for provisioning (e.g., `[eu-west-1, us-west4, westeurope]`)", "type": "array", "items": { "type": "string" } }, "availability_zones": { "title": "Availability Zones", "description": "The availability zones to consider for provisioning (e.g., `[eu-west-1a, us-west4-a]`)", "type": "array", "items": { "type": "string" } }, "instance_types": { "title": "Instance Types", "description": "The cloud-specific instance types to consider for provisioning (e.g., `[p3.8xlarge, n1-standard-4]`)", "type": "array", "items": { "type": "string" } }, "reservation": { "title": "Reservation", "description": "The existing reservation to use for instance provisioning. Supports AWS Capacity Reservations, AWS Capacity Blocks, and GCP reservations", "type": "string" }, "spot_policy": { "description": "The policy for provisioning spot or on-demand instances: `spot`, `on-demand`, `auto`. Defaults to `on-demand`", "allOf": [ { "$ref": "#/definitions/SpotPolicy" } ] }, "retry": { "title": "Retry", "description": "The policy for resubmitting the run. Defaults to `false`", "anyOf": [ { "$ref": "#/definitions/ProfileRetryRequest" }, { "type": "boolean" } ] }, "max_duration": { "title": "Max Duration", "description": "The maximum duration of a run (e.g., `2h`, `1d`, etc) in a running state, excluding provisioning and pulling. After it elapses, the run is automatically stopped. Use `off` for unlimited duration. Defaults to `off`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "stop_duration": { "title": "Stop Duration", "description": "The maximum duration of a run graceful stopping. After it elapses, the run is automatically forced stopped. This includes force detaching volumes used by the run. Use `off` for unlimited duration. Defaults to `5m`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "max_price": { "title": "Max Price", "description": "The maximum instance price per hour, in dollars", "exclusiveMinimum": 0.0, "type": "number" }, "creation_policy": { "description": "The policy for using instances from fleets: `reuse`, `reuse-or-create`. Defaults to `reuse-or-create`", "allOf": [ { "$ref": "#/definitions/CreationPolicy" } ] }, "idle_duration": { "title": "Idle Duration", "description": "Time to wait before terminating idle instances. When the run reuses an existing fleet instance, the fleet's `idle_duration` applies. When the run provisions a new instance, the shorter of the fleet's and run's values is used. Defaults to `5m` for runs and `3d` for fleets. Use `off` for unlimited duration. Only applied for VM-based backends", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "utilization_policy": { "title": "Utilization Policy", "description": "Run termination policy based on utilization", "allOf": [ { "$ref": "#/definitions/UtilizationPolicyRequest" } ] }, "startup_order": { "description": "The order in which master and workers jobs are started: `any`, `master-first`, `workers-first`. Defaults to `any`", "allOf": [ { "$ref": "#/definitions/StartupOrder" } ] }, "stop_criteria": { "description": "The criteria determining when a multi-node run should be considered finished: `all-done`, `master-done`. Defaults to `all-done`", "allOf": [ { "$ref": "#/definitions/StopCriteria" } ] }, "schedule": { "title": "Schedule", "description": "The schedule for starting the run at specified time", "allOf": [ { "$ref": "#/definitions/ScheduleRequest" } ] }, "fleets": { "title": "Fleets", "description": "The fleets considered for reuse. For fleets owned by the current project, specify fleet names. For imported fleets, specify `/`", "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/EntityReferenceRequest" }, { "type": "string" } ] } }, "tags": { "title": "Tags", "description": "The custom tags to associate with the resource. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "additionalProperties": false }, "TaskConfigurationRequest": { "title": "TaskConfigurationRequest", "type": "object", "properties": { "nodes": { "title": "Nodes", "description": "Number of nodes", "default": 1, "minimum": 1, "type": "integer" }, "ports": { "title": "Ports", "description": "Port numbers/mapping to expose", "default": [], "type": "array", "items": { "anyOf": [ { "type": "integer", "exclusiveMinimum": 0, "maximum": 65536 }, { "type": "string", "pattern": "^(?:[0-9]+|\\*):[0-9]+$" }, { "$ref": "#/definitions/PortMappingRequest" } ] } }, "commands": { "title": "Commands", "description": "The shell commands to run", "default": [], "type": "array", "items": { "type": "string" } }, "type": { "title": "Type", "default": "task", "enum": [ "task" ], "type": "string" }, "name": { "title": "Name", "description": "The run name. If not specified, a random name is generated", "type": "string" }, "image": { "title": "Image", "description": "The name of the Docker image to run", "type": "string" }, "user": { "title": "User", "description": "The user inside the container, `user_name_or_id[:group_name_or_id]` (e.g., `ubuntu`, `1000:1000`). Defaults to the default user from the `image`", "type": "string" }, "privileged": { "title": "Privileged", "description": "Run the container in privileged mode", "default": false, "type": "boolean" }, "entrypoint": { "title": "Entrypoint", "description": "The Docker entrypoint", "type": "string" }, "working_dir": { "title": "Working Dir", "description": "The absolute path to the working directory inside the container. Defaults to the `image`'s default working directory", "type": "string" }, "home_dir": { "title": "Home Dir", "default": "/root", "type": "string" }, "registry_auth": { "title": "Registry Auth", "description": "Credentials for pulling a private Docker image", "allOf": [ { "$ref": "#/definitions/RegistryAuthRequest" } ] }, "python": { "description": "The major version of Python. Mutually exclusive with `image` and `docker`", "allOf": [ { "$ref": "#/definitions/PythonVersion" } ] }, "nvcc": { "title": "Nvcc", "description": "Use image with NVIDIA CUDA Compiler (NVCC) included. Mutually exclusive with `image` and `docker`", "type": "boolean" }, "single_branch": { "title": "Single Branch", "description": "Whether to clone and track only the current branch or all remote branches. Relevant only when using remote Git repos. Defaults to `false` for dev environments and to `true` for tasks and services", "type": "boolean" }, "env": { "title": "Env", "description": "The mapping or the list of environment variables", "default": { "__root__": {} }, "allOf": [ { "$ref": "#/definitions/Env" } ] }, "shell": { "title": "Shell", "description": "The shell used to run commands. Allowed values are `sh`, `bash`, or an absolute path, e.g., `/usr/bin/zsh`. Defaults to `/bin/sh` if the `image` is specified, `/bin/bash` otherwise", "type": "string" }, "resources": { "title": "Resources", "description": "The resources requirements to run the configuration", "default": { "cpu": { "min": 2, "max": null }, "memory": { "min": 8.0, "max": null }, "shm_size": null, "gpu": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "disk": { "size": { "min": 100.0, "max": null } } }, "allOf": [ { "$ref": "#/definitions/ResourcesSpecRequest" } ] }, "priority": { "title": "Priority", "description": "The priority of the run, an integer between `0` and `100`. `dstack` tries to provision runs with higher priority first. Defaults to `0`", "minimum": 0, "maximum": 100, "type": "integer" }, "volumes": { "title": "Volumes", "description": "The volumes mount points", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/VolumeMountPointRequest" }, { "$ref": "#/definitions/InstanceMountPointRequest" }, { "type": "string" } ] } }, "docker": { "title": "Docker", "description": "Use Docker inside the container. Mutually exclusive with `image`, `python`, and `nvcc`. Overrides `privileged`", "type": "boolean" }, "repos": { "title": "Repos", "description": "The list of Git repos", "default": [], "type": "array", "items": { "$ref": "#/definitions/RepoSpecRequest" } }, "files": { "title": "Files", "description": "The local to container file path mappings", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/FilePathMappingRequest" }, { "type": "string" } ] } }, "setup": { "title": "Setup", "default": [], "type": "array", "items": { "type": "string" } }, "backends": { "description": "The backends to consider for provisioning (e.g., `[aws, gcp]`)", "type": "array", "items": { "$ref": "#/definitions/BackendType" } }, "regions": { "title": "Regions", "description": "The regions to consider for provisioning (e.g., `[eu-west-1, us-west4, westeurope]`)", "type": "array", "items": { "type": "string" } }, "availability_zones": { "title": "Availability Zones", "description": "The availability zones to consider for provisioning (e.g., `[eu-west-1a, us-west4-a]`)", "type": "array", "items": { "type": "string" } }, "instance_types": { "title": "Instance Types", "description": "The cloud-specific instance types to consider for provisioning (e.g., `[p3.8xlarge, n1-standard-4]`)", "type": "array", "items": { "type": "string" } }, "reservation": { "title": "Reservation", "description": "The existing reservation to use for instance provisioning. Supports AWS Capacity Reservations, AWS Capacity Blocks, and GCP reservations", "type": "string" }, "spot_policy": { "description": "The policy for provisioning spot or on-demand instances: `spot`, `on-demand`, `auto`. Defaults to `on-demand`", "allOf": [ { "$ref": "#/definitions/SpotPolicy" } ] }, "retry": { "title": "Retry", "description": "The policy for resubmitting the run. Defaults to `false`", "anyOf": [ { "$ref": "#/definitions/ProfileRetryRequest" }, { "type": "boolean" } ] }, "max_duration": { "title": "Max Duration", "description": "The maximum duration of a run (e.g., `2h`, `1d`, etc) in a running state, excluding provisioning and pulling. After it elapses, the run is automatically stopped. Use `off` for unlimited duration. Defaults to `off`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "stop_duration": { "title": "Stop Duration", "description": "The maximum duration of a run graceful stopping. After it elapses, the run is automatically forced stopped. This includes force detaching volumes used by the run. Use `off` for unlimited duration. Defaults to `5m`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "max_price": { "title": "Max Price", "description": "The maximum instance price per hour, in dollars", "exclusiveMinimum": 0.0, "type": "number" }, "creation_policy": { "description": "The policy for using instances from fleets: `reuse`, `reuse-or-create`. Defaults to `reuse-or-create`", "allOf": [ { "$ref": "#/definitions/CreationPolicy" } ] }, "idle_duration": { "title": "Idle Duration", "description": "Time to wait before terminating idle instances. When the run reuses an existing fleet instance, the fleet's `idle_duration` applies. When the run provisions a new instance, the shorter of the fleet's and run's values is used. Defaults to `5m` for runs and `3d` for fleets. Use `off` for unlimited duration. Only applied for VM-based backends", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "utilization_policy": { "title": "Utilization Policy", "description": "Run termination policy based on utilization", "allOf": [ { "$ref": "#/definitions/UtilizationPolicyRequest" } ] }, "startup_order": { "description": "The order in which master and workers jobs are started: `any`, `master-first`, `workers-first`. Defaults to `any`", "allOf": [ { "$ref": "#/definitions/StartupOrder" } ] }, "stop_criteria": { "description": "The criteria determining when a multi-node run should be considered finished: `all-done`, `master-done`. Defaults to `all-done`", "allOf": [ { "$ref": "#/definitions/StopCriteria" } ] }, "schedule": { "title": "Schedule", "description": "The schedule for starting the run at specified time", "allOf": [ { "$ref": "#/definitions/ScheduleRequest" } ] }, "fleets": { "title": "Fleets", "description": "The fleets considered for reuse. For fleets owned by the current project, specify fleet names. For imported fleets, specify `/`", "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/EntityReferenceRequest" }, { "type": "string" } ] } }, "tags": { "title": "Tags", "description": "The custom tags to associate with the resource. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "additionalProperties": false }, "TGIChatModelRequest": { "title": "TGIChatModelRequest", "description": "Mapping of the model for the OpenAI-compatible endpoint.\n\nAttributes:\n type (str): The type of the model, e.g. \"chat\"\n name (str): The name of the model. This name will be used both to load model configuration from the HuggingFace Hub and in the OpenAI-compatible endpoint.\n format (str): The format of the model, e.g. \"tgi\" if the model is served with HuggingFace's Text Generation Inference.\n chat_template (Optional[str]): The custom prompt template for the model. If not specified, the default prompt template from the HuggingFace Hub configuration will be used.\n eos_token (Optional[str]): The custom end of sentence token. If not specified, the default end of sentence token from the HuggingFace Hub configuration will be used.", "type": "object", "properties": { "type": { "title": "Type", "description": "The type of the model", "default": "chat", "enum": [ "chat" ], "type": "string" }, "name": { "title": "Name", "description": "The name of the model", "type": "string" }, "format": { "title": "Format", "description": "The serving format. Must be set to `tgi`", "enum": [ "tgi" ], "type": "string" }, "chat_template": { "title": "Chat Template", "description": "The custom prompt template for the model. If not specified, the default prompt template from the HuggingFace Hub configuration will be used", "type": "string" }, "eos_token": { "title": "Eos Token", "description": "The custom end of sentence token. If not specified, the default end of sentence token from the HuggingFace Hub configuration will be used", "type": "string" } }, "required": [ "name", "format" ], "additionalProperties": false }, "OpenAIChatModelRequest": { "title": "OpenAIChatModelRequest", "description": "Mapping of the model for the OpenAI-compatible endpoint.\n\nAttributes:\n type (str): The type of the model, e.g. \"chat\"\n name (str): The name of the model. This name will be used both to load model configuration from the HuggingFace Hub and in the OpenAI-compatible endpoint.\n format (str): The format of the model, i.e. \"openai\".\n prefix (str): The `base_url` prefix: `http://hostname/{prefix}/chat/completions`. Defaults to `/v1`.", "type": "object", "properties": { "type": { "title": "Type", "description": "The type of the model", "default": "chat", "enum": [ "chat" ], "type": "string" }, "name": { "title": "Name", "description": "The name of the model", "type": "string" }, "format": { "title": "Format", "description": "The serving format. Must be set to `openai`", "enum": [ "openai" ], "type": "string" }, "prefix": { "title": "Prefix", "description": "The `base_url` prefix (after hostname)", "default": "/v1", "type": "string" } }, "required": [ "name", "format" ], "additionalProperties": false }, "ScalingSpecRequest": { "title": "ScalingSpecRequest", "type": "object", "properties": { "metric": { "title": "Metric", "description": "The target metric to track. Currently, the only supported value is `rps` (meaning requests per second)", "enum": [ "rps" ], "type": "string" }, "target": { "title": "Target", "description": "The target value of the metric. The number of replicas is calculated based on this number and automatically adjusts (scales up or down) as this metric changes", "exclusiveMinimum": 0, "type": "number" }, "scale_up_delay": { "title": "Scale Up Delay", "description": "The delay in seconds before scaling up", "default": 300, "type": "integer" }, "scale_down_delay": { "title": "Scale Down Delay", "description": "The delay in seconds before scaling down", "default": 600, "type": "integer" } }, "required": [ "metric", "target" ], "additionalProperties": false }, "IPAddressPartitioningKeyRequest": { "title": "IPAddressPartitioningKeyRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "Partitioning type", "default": "ip_address", "enum": [ "ip_address" ], "type": "string" } }, "additionalProperties": false }, "HeaderPartitioningKeyRequest": { "title": "HeaderPartitioningKeyRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "Partitioning type", "default": "header", "enum": [ "header" ], "type": "string" }, "header": { "title": "Header", "description": "Name of the header to use for partitioning", "maxLength": 500, "pattern": "^[a-zA-Z0-9-_]+$", "type": "string" } }, "required": [ "header" ], "additionalProperties": false }, "RateLimitRequest": { "title": "RateLimitRequest", "type": "object", "properties": { "prefix": { "title": "Prefix", "description": "URL path prefix to which this limit is applied. If an incoming request matches several prefixes, the longest prefix is applied", "default": "/", "maxLength": 4094, "pattern": "^/[^\\s\\\\{}]*$", "type": "string" }, "key": { "title": "Key", "description": "The partitioning key. Each incoming request belongs to a partition and rate limits are applied per partition. Defaults to partitioning by client IP address", "default": { "type": "ip_address" }, "discriminator": { "propertyName": "type", "mapping": { "ip_address": "#/definitions/IPAddressPartitioningKeyRequest", "header": "#/definitions/HeaderPartitioningKeyRequest" } }, "oneOf": [ { "$ref": "#/definitions/IPAddressPartitioningKeyRequest" }, { "$ref": "#/definitions/HeaderPartitioningKeyRequest" } ] }, "rps": { "title": "Rps", "description": "Max allowed number of requests per second. Requests are tracked at millisecond granularity. For example, `rps: 10` means at most 1 request per 100ms", "minimum": 0.016666666666666666, "maximum": 153722867280912930, "type": "number" }, "burst": { "title": "Burst", "description": "Max number of requests that can be passed to the service ahead of the rate limit", "default": 0, "minimum": 0, "maximum": 9223372036854775807, "type": "integer" } }, "required": [ "rps" ], "additionalProperties": false }, "HTTPHeaderSpecRequest": { "title": "HTTPHeaderSpecRequest", "type": "object", "properties": { "name": { "title": "Name", "description": "The name of the HTTP header", "maxLength": 256, "minLength": 1, "type": "string" }, "value": { "title": "Value", "description": "The value of the HTTP header", "maxLength": 2048, "minLength": 1, "type": "string" } }, "required": [ "name", "value" ], "additionalProperties": false }, "ProbeConfigRequest": { "title": "ProbeConfigRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "The probe type. Must be `http`", "enum": [ "http" ], "type": "string" }, "url": { "title": "Url", "description": "The URL to request. Defaults to `/`", "type": "string" }, "method": { "title": "Method", "description": "The HTTP method to use for the probe (e.g., `get`, `post`, etc.). Defaults to `get`", "enum": [ "get", "post", "put", "delete", "patch", "head" ], "type": "string" }, "headers": { "title": "Headers", "description": "A list of HTTP headers to include in the request", "default": [], "maxItems": 16, "type": "array", "items": { "$ref": "#/definitions/HTTPHeaderSpecRequest" } }, "body": { "title": "Body", "description": "The HTTP request body to send with the probe", "maxLength": 2048, "minLength": 1, "type": "string" }, "timeout": { "title": "Timeout", "description": "Maximum amount of time the HTTP request is allowed to take. Defaults to `10s`", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "interval": { "title": "Interval", "description": "Minimum amount of time between the end of one probe execution and the start of the next. Defaults to `15s`", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "ready_after": { "title": "Ready After", "description": "The number of consecutive successful probe executions required for the replica to be considered ready. Used during rolling deployments. Defaults to `1`", "minimum": 1, "type": "integer" }, "until_ready": { "title": "Until Ready", "description": "If `true`, the probe will stop being executed as soon as it reaches the `ready_after` threshold of successful executions. Defaults to `false`", "type": "boolean" } }, "required": [ "type" ], "additionalProperties": false }, "ReplicaGroupRequest": { "title": "ReplicaGroupRequest", "type": "object", "properties": { "name": { "title": "Name", "description": "The name of the replica group. If not provided, defaults to '0', '1', etc. based on position.", "type": "string" }, "count": { "title": "Count", "description": "The number of replicas. Can be a number (e.g. `2`) or a range (`0..4` or `1..8`). If it's a range, the `scaling` property is required", "allOf": [ { "$ref": "#/definitions/Range_int_" } ] }, "scaling": { "title": "Scaling", "description": "The auto-scaling rules. Required if `count` is set to a range", "allOf": [ { "$ref": "#/definitions/ScalingSpecRequest" } ] }, "resources": { "title": "Resources", "description": "The resources requirements for replicas in this group", "default": { "cpu": { "min": 2, "max": null }, "memory": { "min": 8.0, "max": null }, "shm_size": null, "gpu": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "disk": { "size": { "min": 100.0, "max": null } } }, "allOf": [ { "$ref": "#/definitions/ResourcesSpecRequest" } ] }, "commands": { "title": "Commands", "description": "The shell commands to run for replicas in this group", "default": [], "type": "array", "items": { "type": "string" } } }, "required": [ "count" ], "additionalProperties": false }, "SGLangServiceRouterConfigRequest": { "title": "SGLangServiceRouterConfigRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "The router type", "default": "sglang", "enum": [ "sglang" ], "type": "string" }, "policy": { "title": "Policy", "description": "The routing policy. Options: `random`, `round_robin`, `cache_aware`, `power_of_two`", "default": "cache_aware", "enum": [ "random", "round_robin", "cache_aware", "power_of_two" ], "type": "string" }, "pd_disaggregation": { "title": "Pd Disaggregation", "description": "Enable PD disaggregation mode for the SGLang router", "default": false, "type": "boolean" } }, "additionalProperties": false }, "ServiceConfigurationRequest": { "title": "ServiceConfigurationRequest", "type": "object", "properties": { "port": { "title": "Port", "description": "The port the application listens on", "anyOf": [ { "type": "integer", "exclusiveMinimum": 0, "maximum": 65536 }, { "type": "string", "pattern": "^[0-9]+:[0-9]+$" }, { "$ref": "#/definitions/PortMappingRequest" } ] }, "gateway": { "title": "Gateway", "description": "The name of the gateway. Specify boolean `false` to run without a gateway. Specify boolean `true` to run with the default gateway. Omit to run with the default gateway if there is one, or without a gateway otherwise", "anyOf": [ { "type": "boolean" }, { "type": "string" } ] }, "strip_prefix": { "title": "Strip Prefix", "description": "Strip the `/proxy/services///` path prefix when forwarding requests to the service. Only takes effect when running the service without a gateway", "default": true, "type": "boolean" }, "model": { "title": "Model", "description": "Mapping of the model for the OpenAI-compatible endpoint provided by `dstack`. Can be a full model format definition or just a model name. If it's a name, the service is expected to expose an OpenAI-compatible API at the `/v1` path", "anyOf": [ { "$ref": "#/definitions/TGIChatModelRequest" }, { "$ref": "#/definitions/OpenAIChatModelRequest" }, { "type": "string" } ] }, "https": { "title": "Https", "description": "Enable HTTPS if running with a gateway. Set to `auto` to determine automatically based on gateway configuration. Defaults to `true`", "anyOf": [ { "type": "boolean" }, { "enum": [ "auto" ], "type": "string" } ] }, "auth": { "title": "Auth", "description": "Enable the authorization", "default": true, "type": "boolean" }, "scaling": { "title": "Scaling", "description": "The auto-scaling rules. Required if `replicas` is set to a range", "allOf": [ { "$ref": "#/definitions/ScalingSpecRequest" } ] }, "rate_limits": { "title": "Rate Limits", "description": "Rate limiting rules", "default": [], "type": "array", "items": { "$ref": "#/definitions/RateLimitRequest" } }, "probes": { "title": "Probes", "description": "The list of probes to determine service health. If `model` is set, defaults to a `/v1/chat/completions` probe. Set explicitly to override", "type": "array", "items": { "$ref": "#/definitions/ProbeConfigRequest" } }, "replicas": { "title": "Replicas", "description": "The number of replicas or a list of replica groups. Can be an integer (e.g., `2`), a range (e.g., `0..4`), or a list of replica groups. Each replica group defines replicas with shared configuration (commands, resources, scaling). When `replicas` is a list of replica groups, top-level `scaling`, `commands`, and `resources` are not allowed and must be specified in each replica group instead. ", "anyOf": [ { "type": "array", "items": { "$ref": "#/definitions/ReplicaGroupRequest" } }, { "$ref": "#/definitions/Range_int_" }, { "type": "integer" }, { "type": "string" } ] }, "router": { "title": "Router", "description": "Router configuration for the service. Requires a gateway with matching router enabled. ", "allOf": [ { "$ref": "#/definitions/SGLangServiceRouterConfigRequest" } ] }, "commands": { "title": "Commands", "description": "The shell commands to run", "default": [], "type": "array", "items": { "type": "string" } }, "type": { "title": "Type", "default": "service", "enum": [ "service" ], "type": "string" }, "name": { "title": "Name", "description": "The run name. If not specified, a random name is generated", "type": "string" }, "image": { "title": "Image", "description": "The name of the Docker image to run", "type": "string" }, "user": { "title": "User", "description": "The user inside the container, `user_name_or_id[:group_name_or_id]` (e.g., `ubuntu`, `1000:1000`). Defaults to the default user from the `image`", "type": "string" }, "privileged": { "title": "Privileged", "description": "Run the container in privileged mode", "default": false, "type": "boolean" }, "entrypoint": { "title": "Entrypoint", "description": "The Docker entrypoint", "type": "string" }, "working_dir": { "title": "Working Dir", "description": "The absolute path to the working directory inside the container. Defaults to the `image`'s default working directory", "type": "string" }, "home_dir": { "title": "Home Dir", "default": "/root", "type": "string" }, "registry_auth": { "title": "Registry Auth", "description": "Credentials for pulling a private Docker image", "allOf": [ { "$ref": "#/definitions/RegistryAuthRequest" } ] }, "python": { "description": "The major version of Python. Mutually exclusive with `image` and `docker`", "allOf": [ { "$ref": "#/definitions/PythonVersion" } ] }, "nvcc": { "title": "Nvcc", "description": "Use image with NVIDIA CUDA Compiler (NVCC) included. Mutually exclusive with `image` and `docker`", "type": "boolean" }, "single_branch": { "title": "Single Branch", "description": "Whether to clone and track only the current branch or all remote branches. Relevant only when using remote Git repos. Defaults to `false` for dev environments and to `true` for tasks and services", "type": "boolean" }, "env": { "title": "Env", "description": "The mapping or the list of environment variables", "default": { "__root__": {} }, "allOf": [ { "$ref": "#/definitions/Env" } ] }, "shell": { "title": "Shell", "description": "The shell used to run commands. Allowed values are `sh`, `bash`, or an absolute path, e.g., `/usr/bin/zsh`. Defaults to `/bin/sh` if the `image` is specified, `/bin/bash` otherwise", "type": "string" }, "resources": { "title": "Resources", "description": "The resources requirements to run the configuration", "default": { "cpu": { "min": 2, "max": null }, "memory": { "min": 8.0, "max": null }, "shm_size": null, "gpu": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "disk": { "size": { "min": 100.0, "max": null } } }, "allOf": [ { "$ref": "#/definitions/ResourcesSpecRequest" } ] }, "priority": { "title": "Priority", "description": "The priority of the run, an integer between `0` and `100`. `dstack` tries to provision runs with higher priority first. Defaults to `0`", "minimum": 0, "maximum": 100, "type": "integer" }, "volumes": { "title": "Volumes", "description": "The volumes mount points", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/VolumeMountPointRequest" }, { "$ref": "#/definitions/InstanceMountPointRequest" }, { "type": "string" } ] } }, "docker": { "title": "Docker", "description": "Use Docker inside the container. Mutually exclusive with `image`, `python`, and `nvcc`. Overrides `privileged`", "type": "boolean" }, "repos": { "title": "Repos", "description": "The list of Git repos", "default": [], "type": "array", "items": { "$ref": "#/definitions/RepoSpecRequest" } }, "files": { "title": "Files", "description": "The local to container file path mappings", "default": [], "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/FilePathMappingRequest" }, { "type": "string" } ] } }, "setup": { "title": "Setup", "default": [], "type": "array", "items": { "type": "string" } }, "backends": { "description": "The backends to consider for provisioning (e.g., `[aws, gcp]`)", "type": "array", "items": { "$ref": "#/definitions/BackendType" } }, "regions": { "title": "Regions", "description": "The regions to consider for provisioning (e.g., `[eu-west-1, us-west4, westeurope]`)", "type": "array", "items": { "type": "string" } }, "availability_zones": { "title": "Availability Zones", "description": "The availability zones to consider for provisioning (e.g., `[eu-west-1a, us-west4-a]`)", "type": "array", "items": { "type": "string" } }, "instance_types": { "title": "Instance Types", "description": "The cloud-specific instance types to consider for provisioning (e.g., `[p3.8xlarge, n1-standard-4]`)", "type": "array", "items": { "type": "string" } }, "reservation": { "title": "Reservation", "description": "The existing reservation to use for instance provisioning. Supports AWS Capacity Reservations, AWS Capacity Blocks, and GCP reservations", "type": "string" }, "spot_policy": { "description": "The policy for provisioning spot or on-demand instances: `spot`, `on-demand`, `auto`. Defaults to `on-demand`", "allOf": [ { "$ref": "#/definitions/SpotPolicy" } ] }, "retry": { "title": "Retry", "description": "The policy for resubmitting the run. Defaults to `false`", "anyOf": [ { "$ref": "#/definitions/ProfileRetryRequest" }, { "type": "boolean" } ] }, "max_duration": { "title": "Max Duration", "description": "The maximum duration of a run (e.g., `2h`, `1d`, etc) in a running state, excluding provisioning and pulling. After it elapses, the run is automatically stopped. Use `off` for unlimited duration. Defaults to `off`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "stop_duration": { "title": "Stop Duration", "description": "The maximum duration of a run graceful stopping. After it elapses, the run is automatically forced stopped. This includes force detaching volumes used by the run. Use `off` for unlimited duration. Defaults to `5m`", "anyOf": [ { "enum": [ "off" ], "type": "string" }, { "type": "integer" }, { "type": "boolean" }, { "type": "string" } ] }, "max_price": { "title": "Max Price", "description": "The maximum instance price per hour, in dollars", "exclusiveMinimum": 0.0, "type": "number" }, "creation_policy": { "description": "The policy for using instances from fleets: `reuse`, `reuse-or-create`. Defaults to `reuse-or-create`", "allOf": [ { "$ref": "#/definitions/CreationPolicy" } ] }, "idle_duration": { "title": "Idle Duration", "description": "Time to wait before terminating idle instances. When the run reuses an existing fleet instance, the fleet's `idle_duration` applies. When the run provisions a new instance, the shorter of the fleet's and run's values is used. Defaults to `5m` for runs and `3d` for fleets. Use `off` for unlimited duration. Only applied for VM-based backends", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "utilization_policy": { "title": "Utilization Policy", "description": "Run termination policy based on utilization", "allOf": [ { "$ref": "#/definitions/UtilizationPolicyRequest" } ] }, "startup_order": { "description": "The order in which master and workers jobs are started: `any`, `master-first`, `workers-first`. Defaults to `any`", "allOf": [ { "$ref": "#/definitions/StartupOrder" } ] }, "stop_criteria": { "description": "The criteria determining when a multi-node run should be considered finished: `all-done`, `master-done`. Defaults to `all-done`", "allOf": [ { "$ref": "#/definitions/StopCriteria" } ] }, "schedule": { "title": "Schedule", "description": "The schedule for starting the run at specified time", "allOf": [ { "$ref": "#/definitions/ScheduleRequest" } ] }, "fleets": { "title": "Fleets", "description": "The fleets considered for reuse. For fleets owned by the current project, specify fleet names. For imported fleets, specify `/`", "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/EntityReferenceRequest" }, { "type": "string" } ] } }, "tags": { "title": "Tags", "description": "The custom tags to associate with the resource. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "required": [ "port" ], "additionalProperties": false }, "SSHKeyRequest": { "title": "SSHKeyRequest", "type": "object", "properties": { "public": { "title": "Public", "type": "string" }, "private": { "title": "Private", "type": "string" } }, "required": [ "public" ], "additionalProperties": false }, "SSHProxyParamsRequest": { "title": "SSHProxyParamsRequest", "type": "object", "properties": { "hostname": { "title": "Hostname", "description": "The IP address or domain of proxy host", "type": "string" }, "port": { "title": "Port", "description": "The SSH port of proxy host", "type": "integer" }, "user": { "title": "User", "description": "The user to log in with for proxy host", "type": "string" }, "identity_file": { "title": "Identity File", "description": "The private key to use for proxy host", "type": "string" }, "ssh_key": { "$ref": "#/definitions/SSHKeyRequest" } }, "required": [ "hostname", "user", "identity_file" ], "additionalProperties": false }, "SSHHostParamsRequest": { "title": "SSHHostParamsRequest", "type": "object", "properties": { "hostname": { "title": "Hostname", "description": "The IP address or domain to connect to", "type": "string" }, "port": { "title": "Port", "description": "The SSH port to connect to for this host", "type": "integer" }, "user": { "title": "User", "description": "The user to log in with for this host", "type": "string" }, "identity_file": { "title": "Identity File", "description": "The private key to use for this host", "type": "string" }, "proxy_jump": { "title": "Proxy Jump", "description": "The SSH proxy configuration for this host", "allOf": [ { "$ref": "#/definitions/SSHProxyParamsRequest" } ] }, "internal_ip": { "title": "Internal Ip", "description": "The internal IP of the host used for communication inside the cluster. If not specified, `dstack` will use the IP address from `network` or from the first found internal network.", "type": "string" }, "ssh_key": { "$ref": "#/definitions/SSHKeyRequest" }, "blocks": { "title": "Blocks", "description": "The amount of blocks to split the instance into, a number or `auto`. `auto` means as many as possible. The number of GPUs and CPUs must be divisible by the number of blocks. Defaults to the top-level `blocks` value.", "anyOf": [ { "enum": [ "auto" ], "type": "string" }, { "type": "integer", "minimum": 1 } ] } }, "required": [ "hostname" ], "additionalProperties": false }, "SSHParamsRequest": { "title": "SSHParamsRequest", "type": "object", "properties": { "user": { "title": "User", "description": "The user to log in with on all hosts", "type": "string" }, "port": { "title": "Port", "description": "The SSH port to connect to", "type": "integer" }, "identity_file": { "title": "Identity File", "description": "The private key to use for all hosts", "type": "string" }, "ssh_key": { "$ref": "#/definitions/SSHKeyRequest" }, "proxy_jump": { "title": "Proxy Jump", "description": "The SSH proxy configuration for all hosts", "allOf": [ { "$ref": "#/definitions/SSHProxyParamsRequest" } ] }, "hosts": { "title": "Hosts", "description": "The per host connection parameters: a hostname or an object that overrides default ssh parameters", "type": "array", "items": { "anyOf": [ { "$ref": "#/definitions/SSHHostParamsRequest" }, { "type": "string" } ] } }, "network": { "title": "Network", "description": "The network address for cluster setup in the format `/`. `dstack` will use IP addresses from this network for communication between hosts. If not specified, `dstack` will use IPs from the first found internal network.", "type": "string" } }, "required": [ "hosts" ], "additionalProperties": false }, "FleetNodesSpecRequest": { "title": "FleetNodesSpecRequest", "type": "object", "properties": { "min": { "title": "Min", "description": "The minimum number of instances to maintain in the fleet", "type": "integer" }, "target": { "title": "Target", "description": "The number of instances to provision on fleet apply. `min` <= `target` <= `max` Defaults to `min`", "type": "integer" }, "max": { "title": "Max", "description": "The maximum number of instances allowed in the fleet. Unlimited if not specified", "type": "integer" } }, "required": [ "min", "target" ], "additionalProperties": false }, "InstanceGroupPlacement": { "title": "InstanceGroupPlacement", "description": "An enumeration.", "enum": [ "any", "cluster" ], "type": "string" }, "FleetConfigurationRequest": { "title": "FleetConfigurationRequest", "type": "object", "properties": { "type": { "title": "Type", "default": "fleet", "enum": [ "fleet" ], "type": "string" }, "name": { "title": "Name", "description": "The fleet name", "type": "string" }, "env": { "title": "Env", "description": "The mapping or the list of environment variables", "default": { "__root__": {} }, "allOf": [ { "$ref": "#/definitions/Env" } ] }, "ssh_config": { "title": "Ssh Config", "description": "The parameters for adding instances via SSH", "allOf": [ { "$ref": "#/definitions/SSHParamsRequest" } ] }, "nodes": { "title": "Nodes", "description": "The number of instances in cloud fleet", "anyOf": [ { "$ref": "#/definitions/FleetNodesSpecRequest" }, { "type": "integer" }, { "type": "string" } ] }, "placement": { "description": "The placement of instances: `any` or `cluster`", "allOf": [ { "$ref": "#/definitions/InstanceGroupPlacement" } ] }, "reservation": { "title": "Reservation", "description": "The existing reservation to use for instance provisioning. Supports AWS Capacity Reservations, AWS Capacity Blocks, and GCP reservations", "type": "string" }, "resources": { "title": "Resources", "description": "The resources requirements", "default": { "cpu": { "min": 2, "max": null }, "memory": { "min": 8.0, "max": null }, "shm_size": null, "gpu": { "vendor": null, "name": null, "count": { "min": 0, "max": null }, "memory": null, "total_memory": null, "compute_capability": null }, "disk": { "size": { "min": 100.0, "max": null } } }, "allOf": [ { "$ref": "#/definitions/ResourcesSpecRequest" } ] }, "blocks": { "title": "Blocks", "description": "The amount of blocks to split the instance into, a number or `auto`. `auto` means as many as possible. The number of GPUs and CPUs must be divisible by the number of blocks. Defaults to `1`, i.e. do not split", "default": 1, "anyOf": [ { "enum": [ "auto" ], "type": "string" }, { "type": "integer", "minimum": 1 } ] }, "backends": { "description": "The backends to consider for provisioning (e.g., `[aws, gcp]`)", "type": "array", "items": { "$ref": "#/definitions/BackendType" } }, "regions": { "title": "Regions", "description": "The regions to consider for provisioning (e.g., `[eu-west-1, us-west4, westeurope]`)", "type": "array", "items": { "type": "string" } }, "availability_zones": { "title": "Availability Zones", "description": "The availability zones to consider for provisioning (e.g., `[eu-west-1a, us-west4-a]`)", "type": "array", "items": { "type": "string" } }, "instance_types": { "title": "Instance Types", "description": "The cloud-specific instance types to consider for provisioning (e.g., `[p3.8xlarge, n1-standard-4]`)", "type": "array", "items": { "type": "string" } }, "spot_policy": { "description": "The policy for provisioning spot or on-demand instances: `spot`, `on-demand`, `auto`. Defaults to `on-demand`", "allOf": [ { "$ref": "#/definitions/SpotPolicy" } ] }, "retry": { "title": "Retry", "description": "The policy for provisioning retry. Defaults to `false`", "anyOf": [ { "$ref": "#/definitions/ProfileRetryRequest" }, { "type": "boolean" } ] }, "max_price": { "title": "Max Price", "description": "The maximum instance price per hour, in dollars", "exclusiveMinimum": 0.0, "type": "number" }, "idle_duration": { "title": "Idle Duration", "description": "Time to wait before terminating idle instances. Instances are not terminated if the fleet is already at `nodes.min`. Defaults to `5m` for runs and `3d` for fleets. Use `off` for unlimited duration", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "tags": { "title": "Tags", "description": "The custom tags to associate with the resource. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "additionalProperties": false }, "SGLangGatewayRouterConfigRequest": { "title": "SGLangGatewayRouterConfigRequest", "description": "Gateway-level router configuration. type and policy only. pd_disaggregation is service-level.", "type": "object", "properties": { "type": { "title": "Type", "description": "The router type enabled on this gateway.", "default": "sglang", "enum": [ "sglang" ], "type": "string" }, "policy": { "title": "Policy", "description": "The routing policy. Deprecated: prefer setting policy in the service's router config. Options: `random`, `round_robin`, `cache_aware`, `power_of_two`", "default": "cache_aware", "enum": [ "random", "round_robin", "cache_aware", "power_of_two" ], "type": "string" } }, "additionalProperties": false }, "LetsEncryptGatewayCertificateRequest": { "title": "LetsEncryptGatewayCertificateRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "Automatic certificates by Let's Encrypt", "default": "lets-encrypt", "enum": [ "lets-encrypt" ], "type": "string" } }, "additionalProperties": false }, "ACMGatewayCertificateRequest": { "title": "ACMGatewayCertificateRequest", "type": "object", "properties": { "type": { "title": "Type", "description": "Certificates by AWS Certificate Manager (ACM)", "default": "acm", "enum": [ "acm" ], "type": "string" }, "arn": { "title": "Arn", "description": "The ARN of the wildcard ACM certificate for the domain", "type": "string" } }, "required": [ "arn" ], "additionalProperties": false }, "GatewayConfigurationRequest": { "title": "GatewayConfigurationRequest", "type": "object", "properties": { "type": { "title": "Type", "default": "gateway", "enum": [ "gateway" ], "type": "string" }, "name": { "title": "Name", "description": "The gateway name", "type": "string" }, "default": { "title": "Default", "description": "Make the gateway default", "default": false, "type": "boolean" }, "backend": { "description": "The gateway backend", "allOf": [ { "$ref": "#/definitions/BackendType" } ] }, "region": { "title": "Region", "description": "The gateway region", "type": "string" }, "instance_type": { "title": "Instance Type", "description": "Backend-specific instance type to use for the gateway instance. Omit to use the backend's default, which is typically a small non-GPU instance", "minLength": 1, "type": "string" }, "router": { "title": "Router", "description": "The router configuration for this gateway. E.g. `{ type: sglang, policy: round_robin }`.", "allOf": [ { "$ref": "#/definitions/SGLangGatewayRouterConfigRequest" } ] }, "domain": { "title": "Domain", "description": "The gateway domain, e.g. `example.com`", "type": "string" }, "public_ip": { "title": "Public Ip", "description": "Allocate public IP for the gateway", "default": true, "type": "boolean" }, "certificate": { "title": "Certificate", "description": "The SSL certificate configuration. Set to `null` to disable. Defaults to `type: lets-encrypt`", "default": { "type": "lets-encrypt" }, "anyOf": [ { "$ref": "#/definitions/LetsEncryptGatewayCertificateRequest" }, { "$ref": "#/definitions/ACMGatewayCertificateRequest" } ] }, "tags": { "title": "Tags", "description": "The custom tags to associate with the gateway. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "required": [ "backend", "region" ], "additionalProperties": false }, "VolumeConfigurationRequest": { "title": "VolumeConfigurationRequest", "type": "object", "properties": { "type": { "title": "Type", "default": "volume", "enum": [ "volume" ], "type": "string" }, "name": { "title": "Name", "description": "The volume name", "type": "string" }, "backend": { "description": "The volume backend", "allOf": [ { "$ref": "#/definitions/BackendType" } ] }, "region": { "title": "Region", "description": "The volume region", "type": "string" }, "availability_zone": { "title": "Availability Zone", "description": "The volume availability zone", "type": "string" }, "size": { "title": "Size", "description": "The volume size. Must be specified when creating new volumes", "type": "number" }, "volume_id": { "title": "Volume Id", "description": "The volume ID. Must be specified when registering external volumes", "type": "string" }, "auto_cleanup_duration": { "title": "Auto Cleanup Duration", "description": "Time to wait after volume is no longer used by any job before deleting it. Defaults to keep the volume indefinitely. Use the value 'off' or -1 to disable auto-cleanup.", "anyOf": [ { "type": "integer" }, { "type": "string" } ] }, "tags": { "title": "Tags", "description": "The custom tags to associate with the volume. The tags are also propagated to the underlying backend resources. If there is a conflict with backend-level tags, does not override them", "type": "object", "additionalProperties": { "type": "string" } } }, "required": [ "backend", "region" ], "additionalProperties": false } } }