Files
headscale-webui/server.py

400 lines
13 KiB
Python
Raw Normal View History

Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
"""Headscale WebUI Flask server."""
import asyncio
import atexit
import datetime
import functools
from multiprocessing import Lock
from typing import Awaitable, Callable, Type, TypeVar
import headscale_api.schema.headscale.v1 as schema
from aiohttp import ClientConnectionError
from apscheduler.schedulers.background import BackgroundScheduler # type: ignore
from betterproto import Message
from flask import Flask, redirect, render_template, url_for
from flask_pydantic.core import validate
from headscale_api.headscale import UnauthorizedError
from markupsafe import Markup
from pydantic import BaseModel, Field
2023-02-21 20:59:05 +09:00
from werkzeug.middleware.proxy_fix import ProxyFix
import renderer
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
from auth import AuthManager
from config import Config, InitCheckError
from headscale import HeadscaleApi
2023-02-21 16:12:42 +09:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
def create_tainted_app(app: Flask, error: InitCheckError) -> Flask:
"""Run tainted version of the Headscale WebUI after encountering an error."""
app.logger.error(
"Encountered error when trying to run initialization checks. Running in "
"tainted mode (only the error page is available). Correct all errors and "
"restart the server."
2023-02-06 04:58:09 +00:00
)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
@app.route("/<path:path>")
def catchall_redirect(path: str): # pylint: disable=unused-argument
return redirect(url_for("error_page"))
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
@app.route("/error")
async def error_page():
return render_template(
"error.html",
error_message=Markup(
"".join(sub_error.format_message() for sub_error in error)
),
)
2023-03-23 19:10:22 +09:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
return app
2023-03-23 19:10:22 +09:00
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
async def create_app() -> Flask:
"""Run Headscale WebUI Flask application.
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
For arguments refer to `Flask.run()` function.
"""
app = Flask(__name__, static_url_path="/static")
app.wsgi_app = ProxyFix( # type: ignore[method-assign]
app.wsgi_app, x_for=1, x_proto=1, x_host=1, x_prefix=1 # type: ignore
2023-02-06 04:58:09 +00:00
)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
try:
# Try to initialize configuration from environment.
config = Config() # type: ignore
with app.app_context():
# Try to create authentication handler (including loading auth config).
auth = AuthManager(config)
# Try to create Headscale API interface.
headscale = HeadscaleApi(config)
# Check health of Headscale API.
if not await headscale.health_check():
raise ClientConnectionError(f"Health check failed on {headscale.base_url}")
except Exception as error: # pylint: disable=broad-exception-caught
# We want to catch broad exception to ensure no errors whatsoever went through
# the environment init.
with app.app_context():
check_error = InitCheckError.from_exception(error)
return create_tainted_app(app, check_error)
app.logger.setLevel(config.log_level)
app.logger.info(
"Headscale-WebUI Version: %s / %s", config.app_version, config.git_branch
)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
app.logger.info("Logger level set to %s.", config.log_level)
app.logger.info("Debug state: %s", config.debug_mode)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
register_pages(app, headscale, auth)
register_api_endpoints(app, headscale, auth)
register_scheduler(app, headscale)
2023-02-15 12:31:22 +09:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
return app
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
def register_pages(app: Flask, headscale: HeadscaleApi, auth: AuthManager):
"""Register user-facing pages."""
config = headscale.app_config
2023-02-22 22:01:18 +09:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
# Convenience short for render_defaults
render_defaults = functools.partial(
renderer.render_defaults, config, auth.oidc_handler
)
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
@app.route("/")
@app.route("/overview")
@auth.require_login
@headscale.key_check_guard
async def overview_page():
return render_template(
"overview.html",
render_page=await renderer.render_overview(headscale),
**render_defaults(),
)
@app.route("/routes", methods=("GET", "POST"))
@auth.require_login
@headscale.key_check_guard
async def routes_page():
return render_template(
"routes.html",
render_page=await renderer.render_routes(headscale),
**render_defaults(),
)
@app.route("/machines", methods=("GET", "POST"))
@auth.require_login
@headscale.key_check_guard
async def machines_page():
return render_template(
"machines.html",
cards=await renderer.render_machines_cards(headscale),
headscale_server=config.hs_server,
inpage_search=renderer.render_search(),
**render_defaults(),
)
@app.route("/users", methods=("GET", "POST"))
@auth.require_login
@headscale.key_check_guard
async def users_page():
return render_template(
"users.html",
cards=await renderer.render_users_cards(headscale),
inpage_search=renderer.render_search(),
)
@app.route("/settings", methods=("GET", "POST"))
@auth.require_login
async def settings_page():
return render_template(
"settings.html",
url=headscale.base_url,
BUILD_DATE=config.build_date,
APP_VERSION=config.app_version,
GIT_REPO_URL=config.git_repo_url,
GIT_COMMIT=config.git_commit,
GIT_BRANCH=config.git_branch,
HS_VERSION=config.hs_version,
**render_defaults(),
)
@app.route("/error")
async def error_page():
"""Error page redirect.
Once we get out of tainted mode, we want to still have this route active so that
users refreshing the page get redirected to the overview page.
"""
return redirect(url_for("overview_page"))
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
@app.route("/logout")
@auth.require_login
@headscale.key_check_guard
async def logout_page():
logout_url = auth.logout()
if logout_url is not None:
return redirect(logout_url)
return redirect(url_for("overview_page"))
2023-03-17 17:21:25 +09:00
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
def register_api_endpoints(app: Flask, headscale: HeadscaleApi, auth: AuthManager):
"""Register Headscale WebUI API endpoints."""
RequestT = TypeVar("RequestT", bound=Message)
ResponseT = TypeVar("ResponseT", bound=Message)
def api_passthrough(
route: str,
request_type: Type[RequestT],
api_method: Callable[[RequestT], Awaitable[ResponseT | str]],
):
"""Passthrough the Headscale API in a concise form.
Arguments:
route -- Flask route to the API endpoint.
request_type -- request model (from headscale_api.schema).
api_method -- backend method to pass through the Flask request.
"""
async def api_passthrough_page(body: RequestT) -> ResponseT | str:
return await api_method(body) # type: ignore
api_passthrough_page.__name__ = route.replace("/", "_")
api_passthrough_page.__annotations__ = {"body": request_type}
return app.route(route, methods=["POST"])(
auth.require_login(
headscale.key_check_guard(
validate()(api_passthrough_page) # type: ignore
)
)
)
class TestKeyRequest(BaseModel):
"""/api/test_key request."""
api_key: str | None = Field(
None, description="API key to test. If None test the current key."
)
@app.route("/api/test_key", methods=("GET", "POST"))
@auth.require_login
@validate()
async def test_key_page(body: TestKeyRequest):
if body.api_key == "":
body.api_key = None
async with headscale.session:
if not await headscale.test_api_key(body.api_key):
return "Unauthenticated", 401
ret = await headscale.renew_api_key()
match ret:
case None:
return "Unauthenticated", 401
case schema.ApiKey():
return ret
case _:
new_key_info = await headscale.get_api_key_info()
if new_key_info is None:
return "Unauthenticated", 401
return new_key_info
class SaveKeyRequest(BaseModel):
"""/api/save_key request."""
api_key: str
@app.route("/api/save_key", methods=["POST"])
@auth.require_login
@validate()
async def save_key_page(body: SaveKeyRequest):
async with headscale.session:
# Test the new API key.
if not await headscale.test_api_key(body.api_key):
return "Key failed testing. Check your key.", 401
try:
headscale.api_key = body.api_key
except OSError:
return "Key did not save properly. Check logs.", 500
key_info = await headscale.get_api_key_info()
if key_info is None:
return "Key saved but error occurred on key info retrieval."
return (
f'Key saved and tested: Key: "{key_info.prefix}", '
f"expiration: {key_info.expiration}"
)
####################################################################################
# Machine API Endpoints
####################################################################################
class UpdateRoutePageRequest(BaseModel):
"""/api/update_route request."""
route_id: int
current_state: bool
@app.route("/api/update_route", methods=["POST"])
@auth.require_login
@validate()
async def update_route_page(body: UpdateRoutePageRequest):
if body.current_state:
return await headscale.disable_route(
schema.DisableRouteRequest(body.route_id)
)
return await headscale.enable_route(schema.EnableRouteRequest(body.route_id))
api_passthrough(
"/api/machine_information",
schema.GetMachineRequest,
headscale.get_machine,
)
api_passthrough(
"/api/delete_machine",
schema.DeleteMachineRequest,
headscale.delete_machine,
)
api_passthrough(
"/api/rename_machine",
schema.RenameMachineRequest,
headscale.rename_machine,
)
api_passthrough(
"/api/move_user",
schema.MoveMachineRequest,
headscale.move_machine,
)
api_passthrough("/api/set_machine_tags", schema.SetTagsRequest, headscale.set_tags)
api_passthrough(
"/api/register_machine",
schema.RegisterMachineRequest,
headscale.register_machine,
)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
####################################################################################
# User API Endpoints
####################################################################################
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
api_passthrough("/api/rename_user", schema.RenameUserRequest, headscale.rename_user)
api_passthrough("/api/add_user", schema.CreateUserRequest, headscale.create_user)
api_passthrough("/api/delete_user", schema.DeleteUserRequest, headscale.delete_user)
api_passthrough("/api/get_users", schema.ListUsersRequest, headscale.list_users)
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
####################################################################################
# Pre-Auth Key API Endpoints
####################################################################################
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
api_passthrough(
"/api/add_preauth_key",
schema.CreatePreAuthKeyRequest,
headscale.create_pre_auth_key,
)
api_passthrough(
"/api/expire_preauth_key",
schema.ExpirePreAuthKeyRequest,
headscale.expire_pre_auth_key,
)
api_passthrough(
"/api/build_preauthkey_table",
schema.ListPreAuthKeysRequest,
functools.partial(renderer.build_preauth_key_table, headscale),
)
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
####################################################################################
# Route API Endpoints
####################################################################################
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
api_passthrough("/api/get_routes", schema.GetRoutesRequest, headscale.get_routes)
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
scheduler_registered: bool = False
scheduler_lock = Lock()
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
def register_scheduler(app: Flask, headscale: HeadscaleApi):
"""Register background scheduler."""
global scheduler_registered # pylint: disable=global-statement
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
with scheduler_lock:
if scheduler_registered:
# For multi-worker set-up, only a single scheduler needs to be enabled.
return
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
scheduler = BackgroundScheduler(
logger=app.logger, timezone=headscale.app_config.timezone
)
scheduler.start() # type: ignore
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
def renew_api_key():
"""Renew API key in a background job."""
app.logger.info("Key renewal schedule triggered...")
try:
if app.ensure_sync(headscale.renew_api_key)() is None: # type: ignore
app.logger.error("Failed to renew the key. Check configuration.")
except UnauthorizedError:
app.logger.error("Current key is invalid. Check configuration.")
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
scheduler.add_job( # type: ignore
renew_api_key,
"interval",
hours=1,
id="renew_api_key",
max_instances=1,
next_run_time=datetime.datetime.now(),
)
2023-02-06 04:58:09 +00:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
atexit.register(scheduler.shutdown) # type: ignore
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
scheduler_registered = True
2023-03-29 13:59:34 +09:00
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
headscale_webui = asyncio.run(create_app())
2023-03-29 13:59:34 +09:00
if __name__ == "__main__":
Complete the major refactor Major part of #73 Unfortunately, it wasn't possible to split it to multiple smaller commits, since the changes touched the entire application substantially. Here is a short list of major changes: 1. Create a separate library (headscale-api), which is used as a convenient abstraction layer providing Pythonic interface with Pydantic. Headscale API is fully asynchronous library, benefitting from improved concurrency for backend requests thus increasing page load speed, e.g., on "Machines" page. 2. Create a common common, validated with flask-pydantic API passthrough layer from GUI to the backend. 3. Move authentication to a separate (auth.py), consolidating the functionality in a single place (with better place for expansion in the future). 4. Move configuration management to a separate module (config.py). Use Pydantic's BaseSettings for reading values from environment, with extensive validation and error reporting. 5. Reduce the number of health checks. - Now, most are performed during server initialization. If any test fails, the server is started in tainted mode, with only the error page exposed (thus reducing the surface of attack in invalid state). - Key checks are implicit in the requests to the backend and guarded by `@headscale.key_check_guard` decorator. - Key renewal is moved to server-side scheduler. 6. Introduce type hints to the level satisfactory for mypy static analysis. Also, enable some other linters in CI and add optional pre-commit hooks. 7. Properly handle some error states. Instead of returning success and handling different responses, if something fails, there is HTTP error code and standard response for it. 8. General formatting, small rewrites for clarity and more idiomatic Python constructs. Signed-off-by: Marek Pikuła <marek.pikula@embevity.com>
2023-04-21 05:26:11 +00:00
headscale_webui.run(host="0.0.0.0")