fix(ci): point KOMODO_HOST to public/internal url and add wait_for_update

.gitea/workflows/deploy.yml

@@ -12,7 +12,7 @@ jobs:
     name: Deploy to Komodo
     runs-on: hp
     env:
-      KOMODO_HOST: ${{ secrets.KOMODO_HOST || 'http://localhost:9120' }}
+      KOMODO_HOST: ${{ secrets.KOMODO_HOST || 'https://komodo.nickmalik.dev' }}
       KOMODO_API_KEY: ${{ secrets.KOMODO_API_KEY }}
       KOMODO_API_SECRET: ${{ secrets.KOMODO_API_SECRET }}
     steps:
@@ -32,16 +32,24 @@ jobs:
           echo "=== Deploying OpenHands Agent Canvas ==="
           python3 scripts/deploy.py
 
-      - name: Verify Health
+      - name: Check Stack Status
         run: |
-          echo "=== Checking health endpoint on port 18000 ==="
-          sleep 10
-          for i in 1 2 3 4 5 6; do
-            if curl -fsSL -o /dev/null "http://127.0.0.1:18000/" 2>/dev/null || curl -fsSL -o /dev/null "http://127.0.0.1:18000/canvas" 2>/dev/null; then
-              echo "OpenHands Agent Canvas is reachable on http://127.0.0.1:18000/ !"
-              exit 0
-            fi
-            echo "Waiting for container startup (attempt $i/6)..."
-            sleep 5
-          done
-          echo "Container is deploying in background. Check status in Komodo Web UI."
+          echo "=== Stack deployment completed. Verifying stack info via Komodo API ==="
+          python3 -c "
+          import urllib.request, json, os
+          host = os.getenv('KOMODO_HOST', 'https://komodo.nickmalik.dev').rstrip('/')
+          key = os.getenv('KOMODO_API_KEY')
+          secret = os.getenv('KOMODO_API_SECRET')
+          if key and secret:
+              req = urllib.request.Request(
+                  f'{host}/read/GetStack',
+                  data=json.dumps({'stack': 'openhands-agent-canvas'}).encode(),
+                  headers={'Content-Type': 'application/json', 'x-api-key': key, 'x-api-secret': secret},
+                  method='POST'
+              )
+              try:
+                  res = json.loads(urllib.request.urlopen(req, timeout=30).read().decode())
+                  print('Stack Status:', res.get('state', 'N/A'))
+              except Exception as e:
+                  print('Note:', e)
+          "

scripts/deploy.py

@@ -2,7 +2,8 @@
 """
 Deploy script for OpenHands Agent Canvas on Komodo Core.
 Reads docker-compose.yml, ensures the Komodo Stack exists on server-tsj08,
-updates the Compose configuration, and triggers deployment via Komodo API.
+updates the Compose configuration, triggers deployment via Komodo API,
+and waits for deployment completion.
 """
 
 import os
@@ -14,6 +15,7 @@ import urllib.error
 
 STACK_NAME = "openhands-agent-canvas"
 DEFAULT_SERVER = "server-tsj08"
+DEFAULT_KOMODO_HOST = "https://komodo.nickmalik.dev"
 
 
 def get_komodo_config():
@@ -28,13 +30,16 @@ def get_komodo_config():
                 import tomllib
                 with open(cfg_path, "rb") as f:
                     cfg = tomllib.load(f)
-                host = host or cfg.get("host", "http://localhost:9120")
+                host = host or cfg.get("host", DEFAULT_KOMODO_HOST)
                 key = key or cfg.get("cli_key", "")
                 secret = secret or cfg.get("cli_secret", "")
             except Exception as e:
                 print(f"[WARN] Failed to parse {cfg_path}: {e}")
 
-    host = host or "http://localhost:9120"
+    # Fallback to public/internal Komodo domain if localhost is passed inside a runner container
+    if not host or "localhost" in host or "127.0.0.1" in host:
+        host = DEFAULT_KOMODO_HOST
+
     if not (key and secret):
         print("::error::KOMODO_API_KEY and KOMODO_API_SECRET must be set.")
         sys.exit(1)
@@ -63,6 +68,39 @@ def api_call(host, key, secret, endpoint, payload):
         raise RuntimeError(f"Komodo API {endpoint} error {e.code}: {err}") from e
 
 
+def wait_for_update(host, key, secret, update_id, max_wait_sec=300):
+    print(f"[INFO] Monitoring deployment update {update_id}...")
+    start_time = time.time()
+    last_status = None
+
+    while time.time() - start_time < max_wait_sec:
+        time.sleep(5)
+        try:
+            updates_data = api_call(host, key, secret, "/read/ListUpdates", {})
+            for u in updates_data.get("updates", []):
+                uid = (u.get("_id") or {}).get("$oid") or u.get("id")
+                if uid == update_id:
+                    status = u.get("status")
+                    if status != last_status:
+                        print(f"  -> Deployment status: {status}")
+                        last_status = status
+                    if status in ("Complete", "Failed", "Cancelled"):
+                        success = u.get("success", False)
+                        for log in (u.get("logs") or [])[-10:]:
+                            print(f"     [log] {log}")
+                        if success:
+                            print(f"[SUCCESS] Deployment completed successfully!")
+                            return True
+                        else:
+                            print(f"::error::Deployment finished with status {status}")
+                            return False
+        except Exception as e:
+            print(f"[WARN] Error polling update status: {e}")
+
+    print(f"[WARN] Deployment is taking longer than {max_wait_sec}s (likely pulling large image). Continuing.")
+    return True
+
+
 def main():
     host, key, secret = get_komodo_config()
     print(f"=== Komodo Target: {host} ===")
@@ -81,7 +119,7 @@ def main():
     elif server_id:
         print(f"[INFO] Target server '{DEFAULT_SERVER}' resolved to id: {server_id}")
     else:
-        print(f"::error::No active servers found in Komodo.")
+        print("::error::No active servers found in Komodo.")
         sys.exit(1)
 
     # 2. Read docker-compose.yml
@@ -130,7 +168,13 @@ def main():
         host, key, secret, "/execute/DeployStack",
         {"stack": STACK_NAME, "services": [], "stop_time": None}
     )
-    print(f"[SUCCESS] Deploy execution dispatched: {deploy_res}")
+    update_id = (deploy_res.get("_id") or {}).get("$oid") or deploy_res.get("id")
+    if update_id:
+        ok = wait_for_update(host, key, secret, update_id, max_wait_sec=300)
+        if not ok:
+            sys.exit(1)
+    else:
+        print(f"[SUCCESS] Deploy execution dispatched: {deploy_res}")
 
 
 if __name__ == "__main__":