Wait for Home Assistant Core to start

If sending the WebSocket message to Home Assistant Core to inform it
about the start of a backup fails, we should reject the backup process
entirely. This prevents taking a backup without Home Assistant Core
being aware of it.

As a side effect, this likely prevents situation where Core does not
learn about the backup progress since we also won't be able to send
WebSocket messages about Job progress updates.

.github/workflows/builder.yml

@@ -320,6 +320,15 @@ jobs:
             exit 1
           fi
 
+      - name: Wait for Home Assistant Core to start
+        run: |
+          echo "Waiting for Home Assistant Core to start"
+          timeout 10m ha supervisor logs -f -n 10000 -b 0 | grep -q "Detect a running Home Assistant instance"
+          if [ "$?" != "0" ]; then
+            echo "Home Assistant Core did not start within 10 minutes"
+            exit 1
+          fi
+
       - name: Create full backup
         id: backup
         run: |

.github/workflows/ci.yaml

@@ -386,7 +386,7 @@ jobs:
             -o console_output_style=count \
             tests
       - name: Upload coverage artifact
-        uses: actions/upload-artifact@330a01c490aca151604b8cf639adc76d48f6c5d4 # v5.0.0
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
         with:
           name: coverage
           path: .coverage
@@ -417,7 +417,7 @@ jobs:
           echo "Failed to restore Python virtual environment from cache"
           exit 1
       - name: Download all coverage artifacts
-        uses: actions/download-artifact@018cc2cf5baa6db3ef3c5f8a56943fffe632ef53 # v6.0.0
+        uses: actions/download-artifact@634f93cb2916e3fdff6788551b99b062d0335ce0 # v5.0.0
         with:
           name: coverage
           path: coverage/

requirements.txt

@@ -17,7 +17,7 @@ faust-cchardet==2.1.19
 gitpython==3.1.45
 jinja2==3.1.6
 log-rate-limit==1.4.2
-orjson==3.11.4
+orjson==3.11.3
 pulsectl==24.12.0
 pyudev==0.24.4
 PyYAML==6.0.3

supervisor/addons/addon.py

@@ -1562,15 +1562,7 @@ class Addon(AddonModel):
                 )
                 break
 
-            # Exponential backoff to spread retries over the throttle window
-            delay = WATCHDOG_RETRY_SECONDS * (1 << max(attempts - 1, 0))
-            _LOGGER.debug(
-                "Watchdog will retry addon %s in %s seconds (attempt %s)",
-                self.name,
-                delay,
-                attempts + 1,
-            )
-            await asyncio.sleep(delay)
+            await asyncio.sleep(WATCHDOG_RETRY_SECONDS)
 
     async def container_state_changed(self, event: DockerContainerStateEvent) -> None:
         """Set addon state from container state."""

supervisor/homeassistant/module.py

@@ -371,6 +371,12 @@ class HomeAssistant(FileConfiguration, CoreSysAttributes):
                 _LOGGER.error,
             ) from err
 
+        if not resp:
+            raise HomeAssistantBackupError(
+                "Preparing backup of Home Assistant Core failed. No response from HA Core.",
+                _LOGGER.error,
+            )
+
         if resp and not resp.get(ATTR_SUCCESS):
             raise HomeAssistantBackupError(
                 f"Preparing backup of Home Assistant Core failed due to: {resp.get(ATTR_ERROR, {}).get(ATTR_MESSAGE, '')}. Check HA Core logs.",

supervisor/homeassistant/websocket.py

@@ -225,6 +225,10 @@ class HomeAssistantWebSocket(CoreSysAttributes):
         # since it makes a new socket connection and we already have one.
         if not connected and not await self.sys_homeassistant.api.check_api_state():
             # No core access, don't try.
+            _LOGGER.debug(
+                "Home Assistant API is not accessible. Not sending WS message: %s",
+                message,
+            )
             return False
 
         if not self._client: