Jul 14 • Code Bugs & Project Issues
Problem deploying with latest adk (1.6.1)
HI All
i am trying to deploy adk_short_bot using latest adk. Here's my deployment script below
this script is being run from the adk_short_bot directory (the one above the agent), however when i deploy on agent engine i keep on getting this error
app.api.factory.utils.UserCodeControlPlaneError: Control plane operation failed due to user code: No module named 'adk_short_bot'
So it seems that somehow when packaging the application something is going wrong
Could anyone assist?
Kind regars
Marco
import os
from vertexai import agent_engines
from vertexai.preview import reasoning_engines
import vertexai
def main():
# 1. Set your Google Cloud project details
PROJECT_ID = "datascience-projects"
LOCATION = "us-central1" # e.g., "us-central1"
STAGING_BUCKET = "gs://adk_short_bot"
# 2. Initialize Vertex AI
vertexai.init(
project=PROJECT_ID,
location=LOCATION,
staging_bucket=STAGING_BUCKET
)
# 3. Import your root agent using absolute import
from adk_short_bot.agent import root_agent
# 4. Create an ADK app with tracing enabled
app_instance = reasoning_engines.AdkApp(
agent=root_agent,
enable_tracing=True
)
# 5. Deploy the agent to Agent Engine
remote_agent = agent_engines.create(
app_instance,
requirements=["google-cloud-aiplatform[adk,agent_engines]"], # minimal requirement
extra_packages=[os.path.join(os.path.dirname(__file__), "adk_short_bot")], # include your agent package
# env_vars={}, # add if you need environment variables
)
print("Agent deployed successfully!")
print("Resource name:", remote_agent.resource_name)
if __name__ == "__main__":
main()
3
7 comments
Problem deploying with latest adk (1.6.1)
skool.com/ai-developer-accelerator
Master AI & software development to build apps and unlock new income streams. Transform ideas into profits. 💡➕🤖➕👨💻🟰💰
Leaderboard (30-day)
1
+53
2
+30
3
+16
4
+14
5
+12
Powered by