API documentation for Glows.ai SDK
Start a new instance using the SDK
| from_image_id required | string Image ID, for example: 6np5yvq2 |
| gpu_name required | string GPU machine name, for example: NVIDIA L40S PCIe |
| custom_meta_key required | string Unique identifier for the instance, other than the instance ID, set by the user |
| instance_category required | string Machine underlying technology type, for example: container, vm |
| region_name required | string Machine region, for example: TW-03, TW-04, etc. |
| remark | string Instance remark name, set by the user |
| envs | string Default environment variables, for example: foo=bar;tik=tok |
| ports | string Custom exported ports, in JSON string format, for example: [{"port":7890,"protocol":2}, {"port":3306,"protocol":1}], where protocol=1 indicates SSH protocol, and protocol=2 indicates HTTP protocol |
| cmd | string Entry command, supported only in container mode. The command must be able to hang; otherwise, the instance will be automatically released once the command finishes. For example: python -m http.server 8080 |
| code | integer Return code, 0 means success, and others (not 0) means failure |
| msg | string Returned message, typically a failure description |
object |
{- "from_image_id": "87p9kq26",
- "gpu_name": "NVIDIA GeForce RTX 4090",
- "custom_meta_key": "order.123456",
- "instance_category": "container",
- "region_name": "TW-03",
- "remark": "this is from sdk",
- "envs": "foo=bar",
- "ports": "[{\"port\":7890,\"protocol\":2}]",
- "cmd": "string"
}{- "code": 0,
- "msg": "",
- "instance": {
- "instanceID": "boyqp4g8",
- "category": "container",
- "status": "SchedQueued",
- "priceCent": 0,
- "hardware": {
- "gpuName": "NVIDIA GeForce RTX 4090",
- "cpuName": "",
- "gpuNum": 0,
- "gpuRamMB": 0,
- "cpuCoresNum": 0,
- "ramMB": 0,
- "diskGB": 0
}, - "baseImage": {
- "imageID": "87p9kq26",
- "imageName": "Ubuntu20.04 CUDA12.2 Python3.9"
}, - "baseSnap": null,
- "sshCred": {
- "user": "root",
- "password": ")Xdk%2XnZr{t6#EL"
}, - "accesses": [
- {
- "protocol": "ssh",
- "listenPort": 22,
- "url": ""
}, - {
- "protocol": "http",
- "listenPort": 8888,
- "url": ""
}, - {
- "protocol": "http",
- "listenPort": 7890,
- "url": ""
}
], - "domain": 2,
- "createTime": {
- "seconds": 1726810944
}, - "cmd": "",
- "envs": "foo=bar",
- "remark": "this is from sdk",
- "metaKey": "order.123456",
- "sourceType": "SourceSDK",
- "regionName": "TW-03"
}
}Retrieve information about a specific instance
| request_id | string Instance ID (instanceID from the response of creating an instance). Either request_id or meta_key is required. |
| meta_key | string The meta_key specified when creating the instance, globally unique. Either request_id or meta_key is required. |
| code | integer Return code, 0 means success, and others (not 0) means failure |
| msg | string Returned message, typically a failure description |
object |
curl --http1.1 \ --request GET \ --url 'https://example.glows.ai/sdk/v1/instance?meta_key=cpu_test.009' \ --header 'Authorization: Bearer YOUR_SDK_TOKEN'
{- "code": 0,
- "msg": "",
- "instance": {
- "instanceID": "boyqp4g8",
- "category": "container",
- "status": "Running",
- "priceCent": 65,
- "hardware": {
- "gpuName": "NVIDIA GeForce RTX 4090",
- "cpuName": "Intel(R) Xeon(R) Platinum 8352V CPU @ 2.10GHz",
- "gpuNum": 1,
- "gpuRamMB": 24564,
- "cpuCoresNum": 10,
- "ramMB": 43008,
- "diskGB": 100
}, - "baseImage": {
- "imageID": "87p9kq26",
- "imageName": "Ubuntu20.04 CUDA12.2 Python3.9"
}, - "baseSnap": null,
- "sshCred": {
- "user": "root",
- "password": ")Xdk%2XnZr{t6#EL"
}, - "accesses": [
- {
- "protocol": "ssh",
- "listenPort": 22,
- "url": "ssh://tw-01.access.glows.ai:15692"
}, - {
- "protocol": "http",
- "listenPort": 8888,
}, - {
- "protocol": "http",
- "listenPort": 7890,
}
], - "domain": 2,
- "createTime": {
- "seconds": 1726810944
}, - "cmd": "",
- "envs": "foo=bar",
- "remark": "this is from sdk",
- "metaKey": "order.123456",
- "sourceType": "SourceSDK",
- "regionName": "TW-03"
}
}Release an instance using request_id or meta_key
| request_id | string Instance ID (instanceID from the response of creating an instance). Either request_id or meta_key is required. |
| meta_key | string The meta_key specified when creating the instance, globally unique. Either request_id or meta_key is required. |
| code | integer Return code, 0 means success, and others (not 0) means failure |
| msg | string Returned message, typically a failure description |
curl --http1.1 \ --request DELETE \ --url 'https://example.glows.ai/sdk/v1/instance?meta_key=cpu_test.009' \ --header 'Authorization: Bearer YOUR_SDK_TOKEN'
{- "code": 0,
- "msg": ""
}Retrieve the statistics of instances
| code | integer Return code, 0 means success, and others (not 0) means failure |
| msg | string Returned message, typically a failure description |
object Instances amount corresponding to each status; status with a statistic of 0 are not displayed |
{- "code": 0,
- "msg": "",
- "stats": {
- "Running": 1
}
}Retrieve a list of instances with pagination and filtering options
| page required | integer Pagination number |
| per_page required | integer Number of pages per pagination |
| ascending | boolean Whether to create in ascending order, default is reverse order |
| statuses | Array of strings Display instances of the specified status |
| code | integer Return code, 0 means success, and others (not 0) means failure |
| msg | string Returned message, typically a failure description |
object | |
Array of objects |
{- "code": 0,
- "msg": "",
- "pagination": {
- "total": 0,
- "numPages": 0,
- "page": 1,
- "perPage": 10
}, - "instances": [ ]
}