fix(api): add /health endpoint for reliable Docker health checks

The health check was using wget --spider on the root path which hits
the swagger catch-all route. BusyBox wget --spider may not handle the
swagger response correctly, causing the API container to be marked
unhealthy and preventing the seed container from starting.

- Add a dedicated /health endpoint registered before middleware
- Update Docker Compose health check to use /health
- Update CI workflow health wait loop to use /health

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: AchoArnold

.github/workflows/integration-test.yml

@@ -37,7 +37,7 @@ jobs:
         run: |
           echo "Waiting for API to be healthy..."
           for i in $(seq 1 40); do
-            if docker compose exec api wget --spider -q http://localhost:8000/ 2>/dev/null; then
+            if docker compose exec api wget -q -O /dev/null http://localhost:8000/health 2>/dev/null; then
               echo "API is healthy!"
               break
             fi

api/pkg/di/container.go

@@ -174,6 +174,11 @@ func (container *Container) App() (app *fiber.App) {
 
 	app = fiber.New()
 
+	// Health check endpoint registered before middleware for reliable Docker health checks
+	app.Get("/health", func(c *fiber.Ctx) error {
+		return c.SendStatus(fiber.StatusOK)
+	})
+
 	if os.Getenv("USE_HTTP_LOGGER") == "true" {
 		app.Use(fiberLogger.New())
 	}

tests/docker-compose.yml

@@ -57,7 +57,8 @@ services:
     environment:
       FIREBASE_CREDENTIALS: "${FIREBASE_CREDENTIALS}"
     healthcheck:
-      test: ["CMD", "wget", "--spider", "-q", "http://localhost:8000/"]
+      test:
+        ["CMD", "wget", "-q", "-O", "/dev/null", "http://localhost:8000/health"]
       interval: 5s
       timeout: 10s
       retries: 20