Files
motia/bitbylaw/steps/vmh/README_SYNC.md
bitbylaw ae1d96f767 Add sync strategy documentation and templates for bidirectional sync between EspoCRM and Advoware
- Introduced SYNC_STRATEGY_ARCHIVE.md detailing the sync process, status values, and flow for updating entities from EspoCRM to Advoware and vice versa.
- Created SYNC_TEMPLATE.md as a guide for implementing new syncs, including field definitions, mapper examples, sync utilities, event handlers, and cron jobs.
- Added README_SYNC.md for the Beteiligte sync event handler, outlining its functionality, event subscriptions, optimizations, error handling, and performance metrics.
2026-02-07 15:54:13 +00:00

170 lines
4.1 KiB
Markdown

# Beteiligte Sync - Event Handler
Event-driven sync handler für bidirektionale Synchronisation von Beteiligten (Stammdaten).
## Subscribes
- `vmh.beteiligte.create` - Neuer Beteiligter in EspoCRM
- `vmh.beteiligte.update` - Änderung in EspoCRM
- `vmh.beteiligte.delete` - Löschung in EspoCRM
- `vmh.beteiligte.sync_check` - Cron-triggered check
## Funktionsweise
### 1. Event empfangen
```json
{
"entity_id": "68e3e7eab49f09adb",
"action": "sync_check",
"source": "cron",
"timestamp": "2026-02-07T16:00:00"
}
```
### 2. Lock acquisition (Redis)
```python
lock_key = f"sync_lock:cbeteiligte:{entity_id}"
acquired = redis.set(lock_key, "locked", nx=True, ex=300)
```
- **Atomar** via Redis `SET NX`
- **TTL**: 5 Minuten (verhindert Deadlocks)
- **Verhindert**: Parallele Syncs derselben Entity
### 3. Routing nach Action
#### CREATE (kein betnr)
```
Map EspoCRM → Advoware
POST /api/v1/advonet/Beteiligte
Response: {betNr: 12345}
Update EspoCRM: betnr=12345, syncStatus=clean (combined!)
```
#### UPDATE (hat betnr)
```
GET /api/v1/advonet/Beteiligte/{betnr}
Timestamp-Vergleich (modifiedAt vs geaendertAm)
├─ espocrm_newer → PUT to Advoware
├─ advoware_newer → PATCH to EspoCRM
├─ conflict → EspoCRM wins + Notification
└─ no_change → Skip
```
### 4. Lock release
```python
await sync_utils.release_sync_lock(
entity_id,
'clean',
extra_fields={'betnr': new_betnr} # Optional: combine operations
)
```
- Updates `syncStatus`, `advowareLastSync`, `syncRetryCount`
- Optional: Merge zusätzliche Felder (betnr, etc.)
- Löscht Redis lock
## Optimierungen
### Redis Distributed Lock
```python
# VORHER: Nicht-atomar (Race Condition möglich)
entity = await get_entity(...)
if entity.syncStatus == 'syncing':
return
await update_entity(..., {'syncStatus': 'syncing'})
# NACHHER: Atomarer Redis lock
acquired = redis.set(lock_key, "locked", nx=True, ex=300)
if not acquired:
return
```
### Combined API Calls
```python
# VORHER: 2 API calls
await release_sync_lock(entity_id, 'clean')
await update_entity(entity_id, {'betnr': new_betnr})
# NACHHER: 1 API call (33% faster)
await release_sync_lock(
entity_id,
'clean',
extra_fields={'betnr': new_betnr}
)
```
### Merge Utility
```python
# Keine Code-Duplikation mehr (3x → 1x)
merged_data = sync_utils.merge_for_advoware_put(
advo_entity,
espo_entity,
mapper
)
```
## Error Handling
### Retriable Errors
- Netzwerk-Timeout → `syncStatus=failed`, retry beim nächsten Cron
- 500 Server Error → `syncStatus=failed`, retry
- Redis unavailable → Fallback zu syncStatus-only lock
### Non-Retriable Errors
- 400 Bad Request → `syncStatus=failed`, keine Auto-Retry
- 404 Not Found → Entity gelöscht, markiere als `deleted_in_advoware`
- 401 Auth Error → `syncStatus=failed`, keine Auto-Retry
### Max Retries
```python
if retry_count >= 5:
syncStatus = 'permanently_failed'
send_notification("Max retries exceeded")
```
## Performance
| Metric | Value |
|--------|-------|
| Latency (CREATE) | ~200ms |
| Latency (UPDATE) | ~250ms |
| API Calls (CREATE) | 2 |
| API Calls (UPDATE) | 2 |
| Lock Timeout | 5 min |
## Dependencies
- [services/espocrm.py](../../services/espocrm.py) - EspoCRM API
- [services/advoware.py](../../services/advoware.py) - Advoware API
- [services/espocrm_mapper.py](../../services/espocrm_mapper.py) - Entity mapper
- [services/beteiligte_sync_utils.py](../../services/beteiligte_sync_utils.py) - Sync utilities
- Redis (localhost:6379, DB 1) - Distributed locking
## Testing
```python
# Test event
event_data = {
'entity_id': '68e3e7eab49f09adb',
'action': 'sync_check',
'source': 'test'
}
await handler(event_data, context)
# Verify
entity = await espocrm.get_entity('CBeteiligte', entity_id)
assert entity['syncStatus'] == 'clean'
assert entity['betnr'] is not None
```
## Siehe auch
- [Beteiligte Sync Docs](../../docs/BETEILIGTE_SYNC.md) - Vollständige Dokumentation
- [Cron Step](beteiligte_sync_cron_step.py) - Findet Entities für Sync
- [Sync Utils](../../services/beteiligte_sync_utils.py) - Helper functions