fix: make health ping initialize pool

Author: Lekssays

src/utils/postgres_db_manager.py

@@ -44,6 +44,9 @@ def ping(self) -> dict:
         import time
         start = time.monotonic()
         try:
+            # The pooled connection is opened lazily; health probes must be able
+            # to establish it before asking for a connection.
+            self._open_pool()
             with self._connect() as conn:
                 conn.execute("SELECT 1")
             return {"ok": True, "latency_ms": round((time.monotonic() - start) * 1000, 2)}

tests/test_pool_store_redis.py

@@ -28,7 +28,7 @@ def _make_manager(monkeypatch, budget_mb, redis_url):
     m = jsm.JoernServerManager(config=load_config(), redis_url=redis_url)
     # Stub the Docker-/network-touching bits so we exercise only coordination.
     m._wait_for_server = lambda port, timeout=120, codebase_hash="": True
-    m._port_healthy = lambda port: True
+    m._port_healthy = lambda port, codebase_hash="": True
     m._plan_server = lambda h: (2, 3072)  # tier-S: 3072 MB reservation each
     return m, fake