# subproc_mgr
## Motivation
Applications that spawn long-running subprocesses need to be able to manage the lifetime of those subprocesses, which is so-far not a truth as the application can exit unexpectedly (e.g., runtime, or buggy high-level software).
This is a simple service that can be deployed globally that proxies subprocess spawning and manages the lifetime. The service will periodically check the lifetime of the subprocesses and kill them if they have exceeded the expected lifetime.
P.S: This approach could be better (easier and cheaper) than a watchdog if you have a consensus conventions on how to spawn subprocesses.
## Usage
1. The service should be started:
```shell
usage: subproc_mgr [-h] [--port PORT] [--period PERIOD] [port] [period]
port: port that the service is deployed on
period: how often the service checks the lifetime of subprocesses
positional arguments:
port <class 'int'>
period <class 'float'>
options:
-h, --help show this help message and exit
--port PORT <class 'int'>
--period PERIOD <class 'float'>
```
2. We send requests (e.g., HTTP requests) to use the service:
```shell
shell> cat usage_example.py
shell> python usage_example.py
```
```python
from typing import TypedDict
import requests, os, shutil
PORT = 5687
class StartProcessRequest(TypedDict):
command: str
args: list[str]
env: dict[str, str]
pid: int
# start python -m http.server
resp = requests.post(
f"http://localhost:{PORT}/spawn",
json=StartProcessRequest(
command=shutil.which("python") or "python",
args=['-m', 'http.server'],
env=dict(os.environ),
pid=os.getpid()
)
)
assert resp.json().get("code") == "success"
# press Ctrl+D, the subprocess exits very sooner
while True:
import time
time.sleep(0.5)
```
3. Press ctrl+D for the terminal that executes `python usage_example.py`. You'll see the following logs from the service.
## Motivation (for the author)
1. Extending VSCode with dynamically added commands/key bindings.
2. Starting a language runtime and managing the lifetime of its parser server.
Raw data
{
"_id": null,
"home_page": "https://github.com/thautwarm/subproc_mgr",
"name": "subproc-mgr",
"maintainer": "",
"docs_url": null,
"requires_python": ">=3.9",
"maintainer_email": "",
"keywords": "subprocess,lifetime",
"author": "thautwarm",
"author_email": "twshere@outlook.com",
"download_url": "",
"platform": "any",
"description": "# subproc_mgr\n\n## Motivation\n\nApplications that spawn long-running subprocesses need to be able to manage the lifetime of those subprocesses, which is so-far not a truth as the application can exit unexpectedly (e.g., runtime, or buggy high-level software).\n\nThis is a simple service that can be deployed globally that proxies subprocess spawning and manages the lifetime. The service will periodically check the lifetime of the subprocesses and kill them if they have exceeded the expected lifetime.\n\nP.S: This approach could be better (easier and cheaper) than a watchdog if you have a consensus conventions on how to spawn subprocesses.\n\n## Usage\n\n1. The service should be started:\n\n ```shell\n usage: subproc_mgr [-h] [--port PORT] [--period PERIOD] [port] [period]\n\n\n port: port that the service is deployed on\n period: how often the service checks the lifetime of subprocesses\n\n\n positional arguments:\n port <class 'int'>\n period <class 'float'>\n\n options:\n -h, --help show this help message and exit\n --port PORT <class 'int'>\n --period PERIOD <class 'float'>\n ```\n\n2. We send requests (e.g., HTTP requests) to use the service:\n\n ```shell\n shell> cat usage_example.py\n shell> python usage_example.py\n ```\n\n ```python\n\n from typing import TypedDict\n import requests, os, shutil\n\n PORT = 5687\n\n class StartProcessRequest(TypedDict):\n command: str\n args: list[str]\n env: dict[str, str]\n pid: int\n\n # start python -m http.server\n resp = requests.post(\n f\"http://localhost:{PORT}/spawn\",\n json=StartProcessRequest(\n command=shutil.which(\"python\") or \"python\",\n args=['-m', 'http.server'],\n env=dict(os.environ),\n pid=os.getpid()\n )\n )\n\n assert resp.json().get(\"code\") == \"success\"\n\n # press Ctrl+D, the subprocess exits very sooner\n while True:\n import time\n time.sleep(0.5)\n ```\n\n3. Press ctrl+D for the terminal that executes `python usage_example.py`. You'll see the following logs from the service.\n\n \n\n## Motivation (for the author)\n\n1. Extending VSCode with dynamically added commands/key bindings.\n2. Starting a language runtime and managing the lifetime of its parser server.\n",
"bugtrack_url": null,
"license": "mit",
"summary": "A service to spawn and manage long-running child processes, following the lifecycle management of \"child processes are attached to parent processes\".",
"version": "0.1.1",
"project_urls": {
"Homepage": "https://github.com/thautwarm/subproc_mgr"
},
"split_keywords": [
"subprocess",
"lifetime"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "5b9f32f0c9482c5118dca6f3ff34de6336618c38fb3c0ed3289886394cb67a08",
"md5": "45ae8c1519a4c311d18262f8f63ed798",
"sha256": "e3217570fbc1e7b933392e17d7765981624ff299adb5118c24614ea8011cdbdb"
},
"downloads": -1,
"filename": "subproc_mgr-0.1.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "45ae8c1519a4c311d18262f8f63ed798",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.9",
"size": 5280,
"upload_time": "2023-05-14T12:26:51",
"upload_time_iso_8601": "2023-05-14T12:26:51.081714Z",
"url": "https://files.pythonhosted.org/packages/5b/9f/32f0c9482c5118dca6f3ff34de6336618c38fb3c0ed3289886394cb67a08/subproc_mgr-0.1.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-05-14 12:26:51",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "thautwarm",
"github_project": "subproc_mgr",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "subproc-mgr"
}
JSON
