fix: make media port test resilient to backend startup races

Poll /health and retry connection-aborted requests inside the test, and
require consecutive CI health checks, to avoid RemoteDisconnected on a
freshly started status-backend container.

Author: friofry

.github/workflows/ci.yml

@@ -25,9 +25,16 @@ jobs:
 
       - name: Wait for backend health
         run: |
+          ok=0
           for i in $(seq 1 60); do
-            if curl -fsS http://127.0.0.1:8080/health; then
-              exit 0
+            if curl -fsS http://127.0.0.1:8080/health >/dev/null; then
+              ok=$((ok + 1))
+              if [ "$ok" -ge 3 ]; then
+                echo "backend healthy ($ok consecutive checks)"
+                exit 0
+              fi
+            else
+              ok=0
             fi
             sleep 5
           done

tests/test_media_server_port.py

@@ -75,12 +75,32 @@ def close(self):
         self._stop = True
 
 
-def post_json(path: str, payload: dict | None = None) -> dict:
-    resp = requests.post(f"{HTTP_BASE}/{path}", json=payload or {}, timeout=60)
-    resp.raise_for_status()
-    if not resp.content:
-        return {}
-    return resp.json()
+def post_json(path: str, payload: dict | None = None, retries: int = 3) -> dict:
+    last_exc: Exception | None = None
+    for attempt in range(retries):
+        try:
+            resp = requests.post(f"{HTTP_BASE}/{path}", json=payload or {}, timeout=60)
+            resp.raise_for_status()
+            if not resp.content:
+                return {}
+            return resp.json()
+        except requests.ConnectionError as exc:
+            last_exc = exc
+            time.sleep(2 * (attempt + 1))
+    raise RuntimeError(f"POST {path} failed after {retries} retries: {last_exc}")
+
+
+def wait_backend_healthy(timeout: float = 180) -> None:
+    deadline = time.monotonic() + timeout
+    last_exc: Exception | None = None
+    while time.monotonic() < deadline:
+        try:
+            if requests.get("http://127.0.0.1:8080/health", timeout=5).status_code == 200:
+                return
+        except requests.RequestException as exc:
+            last_exc = exc
+        time.sleep(2)
+    raise RuntimeError(f"backend not healthy within {timeout}s: {last_exc}")
 
 
 def initialize() -> dict:
@@ -176,6 +196,7 @@ def signals():
 
 
 def test_media_port_persists_across_logout_reinit(signals):
+    wait_backend_healthy()
     logout(signals)
     since = signals.count("mediaserver.started")
     init_data = initialize()