fix: harden Railway startup — fault-tolerant lifespan + DB retry + credential cleanup

- Wrap PostHog/DLQ init in try/except so startup survives missing services
- Delay self-improvement worker 30s to reduce startup load
- Run init_db() for ALL database types (was SQLite-only, skipping PostgreSQL)
- Add 3-attempt retry with backoff in init_db() for Railway DB startup race
- Fix FastAPI deprecation: regex → pattern in intelligence.py
- Remove hardcoded Ultramsg credentials from auto_pipeline.py

https://claude.ai/code/session_01W1rJthWDkasijTdXCfxVHs

salesflow-saas/backend/app/api/v1/intelligence.py

@@ -126,7 +126,7 @@ async def alert_stats(tenant_id: str):
 
 @router.get("/digest", summary="Generate Arabic alert digest")
 async def generate_digest(tenant_id: str, user_id: Optional[str] = None,
-                           period: str = Query(default="daily", regex="^(daily|weekly)$")):
+                           period: str = Query(default="daily", pattern="^(daily|weekly)$")):
     return await get_alert_delivery().generate_digest(tenant_id, user_id, period)
 
 

salesflow-saas/backend/app/database.py

@@ -70,13 +70,24 @@ async def get_db():
 async def init_db():
     import app.models  # noqa: F401 — register all models on Base.metadata before create_all
 
-    async with engine.begin() as conn:
+    for attempt in range(3):
-        if not IS_SQLITE:
+        try:
-            for ext in ["CREATE EXTENSION IF NOT EXISTS vector",
+            async with engine.begin() as conn:
-                        "CREATE EXTENSION IF NOT EXISTS pg_trgm"]:
+                if not IS_SQLITE:
-                try:
+                    for ext in ["CREATE EXTENSION IF NOT EXISTS vector",
-                    await conn.execute(text(ext))
+                                "CREATE EXTENSION IF NOT EXISTS pg_trgm"]:
-                except Exception:
+                        try:
-                    pass
+                            await conn.execute(text(ext))
-        await conn.run_sync(Base.metadata.create_all)
+                        except Exception:
-        print("✅ Database initialized")
+                            pass
+                await conn.run_sync(Base.metadata.create_all)
+                print("✅ Database initialized")
+                return
+        except Exception as e:
+            if attempt < 2:
+                import asyncio
+                print(f"⚠️ DB init attempt {attempt + 1} failed: {e}, retrying in 3s...")
+                await asyncio.sleep(3)
+            else:
+                print(f"❌ DB init failed after 3 attempts: {e}")
+                raise

salesflow-saas/backend/app/main.py

@@ -50,18 +50,22 @@ async def lifespan(app: FastAPI):
     stop_event = asyncio.Event()
 
     async def _self_improvement_worker():
+        await asyncio.sleep(30)
         while not stop_event.is_set():
-            self_improvement_flow.run(
+            try:
-                tenant_id="system_tenant",
+                self_improvement_flow.run(
-                input_state={
+                    tenant_id="system_tenant",
-                    "signals": [],
+                    input_state={
-                    "bottlenecks": [],
+                        "signals": [],
-                    "experiments": [{"name": "always-on-ab-loop"}],
+                        "bottlenecks": [],
-                    "ab_results": {},
+                        "experiments": [{"name": "always-on-ab-loop"}],
-                    "governance_passed": True,
+                        "ab_results": {},
-                    "promoted": True,
+                        "governance_passed": True,
-                },
+                        "promoted": True,
-            )
+                    },
+                )
+            except Exception:
+                pass
             await asyncio.sleep(max(60, settings.SELF_IMPROVEMENT_INTERVAL_SECONDS))
 
     worker_task = asyncio.create_task(_self_improvement_worker())
@@ -72,16 +76,20 @@ async def lifespan(app: FastAPI):
     print(f"   LLM Primary: {settings.LLM_PRIMARY_PROVIDER}")
     print(f"   LLM Fallback: {settings.LLM_FALLBACK_PROVIDER}")
 
-    # Initialize PostHog
+    try:
-    from app.services.posthog_client import get_posthog
+        from app.services.posthog_client import get_posthog
-    ph = get_posthog()
+        ph = get_posthog()
-    print(f"   PostHog: {'enabled' if ph._enabled else 'disabled (no API key)'}")
+        print(f"   PostHog: {'enabled' if getattr(ph, '_enabled', False) else 'disabled (no API key)'}")
+    except Exception as e:
+        print(f"   PostHog: init failed ({e})")
 
-    # Initialize DLQ
+    try:
-    from app.services.dlq import dlq
+        from app.services.dlq import dlq  # noqa: F841
-    print("   DLQ: initialized")
+        print("   DLQ: initialized")
-    if IS_SQLITE:
+    except Exception as e:
-        await init_db()
+        print(f"   DLQ: init failed ({e})")
+
+    await init_db()
     yield
     # Shutdown
     stop_event.set()

salesflow-saas/backend/app/services/auto_pipeline.py

@@ -115,8 +115,8 @@ class WhatsAppMessenger:
     """Send messages via Ultramsg API."""
 
     def __init__(self):
-        self.instance_id = os.getenv("ULTRAMSG_INSTANCE_ID", "instance168132")
+        self.instance_id = os.getenv("ULTRAMSG_INSTANCE_ID", "")
-        self.token = os.getenv("ULTRAMSG_TOKEN", "7azj2ss74wpg9jwp")
+        self.token = os.getenv("ULTRAMSG_TOKEN", "")
         self.api_base = f"https://api.ultramsg.com/{self.instance_id}"
 
     async def send_message(self, phone: str, message: str) -> Dict: