refactor(backend): clean up jail service, add error handling service

- Extract jail status/processing to helper functions
- Add error_handling.py service for centralized error handling
- Update config.py with validation and defaults
- Update .env.example with all config options
- Remove obsolete Tasks.md, add Service-Development.md
- Minor fixes across routers and services

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
2026-05-03 17:40:37 +02:00
parent 2df029f7e8
commit 2f9fc8076d
15 changed files with 332 additions and 154 deletions

115
Docs/Service-Development.md Normal file
View File

@@ -0,0 +1,115 @@
# Service Development Guide
How to write and maintain services in BanGUI.
## Error Handling Contracts
Every service method must document which error handling pattern it follows.
This lets callers know what to expect without reading the implementation.
### The Three Patterns
```python
from app.services.error_handling import ABORT_ON_ERROR, RETURN_DEFAULT, PARTIAL_RESULT
```
**ABORT_ON_ERROR** — Raise an exception, let the router convert it to HTTP.
Used for: auth, writes, state changes, any operation where partial success is meaningless.
```python
async def start_jail(socket_path: str, name: str) -> None:
"""Start a stopped fail2ban jail.
Error contract: ABORT_ON_ERROR. Raises JailNotFoundError (404),
JailOperationError (409), Fail2BanConnectionError (503).
"""
...
```
**RETURN_DEFAULT** — Return empty result and log warning. Never raises.
Used for: informational reads (list, get) where infrastructure unavailability
should not block the UI.
```python
async def get_settings(socket_path: str) -> DomainServerSettingsResult:
"""Return current fail2ban server-level settings.
Error contract: RETURN_DEFAULT. Returns DomainServerSettingsResult
with default values if socket is unreachable. Never raises.
"""
...
```
**PARTIAL_RESULT** — Return (result, errors) tuple. Errors collected, not raised.
Used for: batch operations on collections where one item failing does not
invalidate the rest.
```python
# Not yet used in codebase; define as needed for batch operations.
```
### When to Use Which
| Operation type | Pattern |
|---------------|---------|
| Auth / session | ABORT_ON_ERROR |
| Write / state change | ABORT_ON_ERROR |
| Config updates | ABORT_ON_ERROR |
| Single-item read (jail, ban) | ABORT_ON_ERROR |
| Multi-item read (list) | RETURN_DEFAULT |
| Server settings read | RETURN_DEFAULT |
| Batch / parallel fetch | PARTIAL_RESULT |
### Changing Patterns
Switching a method's error contract is a **breaking change**. Update the docstring,
add a changelog entry, and bump the major version if this is a public API.
## Service Structure
Services live in `backend/app/services/`. They contain **no** HTTP/FastAPI concerns.
```
app/services/
ban_service.py # ban/unban, ban history queries
jail_service.py # jail lifecycle, ignore lists
server_service.py # server-level settings
geo_service.py # geolocation
...
error_handling.py # contract definitions
protocols.py # Protocol interfaces for DI
```
## Protocols
Each service has a corresponding protocol in `protocols.py` for dependency injection.
Protocol methods include the error contract in their docstring:
```python
class JailService(Protocol):
async def list_jails(self, socket_path: str) -> DomainJailList:
"""Error contract: ABORT_ON_ERROR."""
...
```
## Router Error Handling
Routers must not catch and silently swallow exceptions from services using
ABORT_ON_ERROR unless they convert to a specific HTTP response.
Let domain exceptions propagate — the global exception handlers handle them.
Exception handler registration (in `main.py`):
- `DomainError` → JSON error response
- `Fail2BanConnectionError` → HTTP 503
- `JailNotFoundError` → HTTP 404
## Logging
Log at the service layer using structlog:
```python
log.info("jail_started", jail=name)
log.warning("socket_unreachable_using_default", socket_path=socket_path)
```
Never log sensitive data (tokens, passwords, IPs in full).