fix: add /healthz endpoint to prevent K8s crash loop

- Added dedicated /healthz route returning 200 OK
- Skip API health check in hooks.server.ts for /healthz path
- Updated K8s liveness/readiness probes to use /healthz instead of /login
- The /login probe was returning 503 when the API was unreachable, causing
  Kubernetes to kill and restart the pod in a loop

kubernetes/deployment.yaml

@@ -35,15 +35,17 @@ spec:
             - containerPort: 3000
           livenessProbe:
             httpGet:
-              path: /login
+              path: /healthz
               port: 3000
             initialDelaySeconds: 5
             periodSeconds: 15
+            failureThreshold: 3
           readinessProbe:
             httpGet:
-              path: /login
+              path: /healthz
               port: 3000
             initialDelaySeconds: 3
             periodSeconds: 5
+            failureThreshold: 2
       imagePullSecrets:
         - name: github-container-registry

src/hooks.server.ts

@@ -101,6 +101,11 @@ function apiUnreachablePage(): Response {
 }
 
 export const handle: Handle = async ({ event, resolve }) => {
+  // Let Kubernetes health probes pass without hitting the external API
+  if (event.url.pathname === "/healthz") {
+    return await resolve(event);
+  }
+
   // Health-check the API before doing anything else.
   // /v1/teapot returns 418 when the API is alive.
   try {

src/routes/healthz/+server.ts

@@ -0,0 +1,8 @@
+import type { RequestHandler } from "./$types";
+
+export const GET: RequestHandler = async () => {
+  return new Response(JSON.stringify({ status: "ok" }), {
+    status: 200,
+    headers: { "Content-Type": "application/json" },
+  });
+};