Monitor hosted agent logs with the Azure Developer CLI

Important

Items marked (preview) in this article are currently in public preview. This preview is provided without a service-level agreement, and we don't recommend it for production workloads. Certain features might not be supported or might have constrained capabilities. For more information, see Supplemental Terms of Use for Microsoft Azure Previews.

Stream and inspect logs from your deployed Microsoft Foundry hosted agent for troubleshooting and observability. You learn how to view console logs, stream in real time, inspect system events, filter by session, and recognize common log patterns.

Prerequisites

• A deployed hosted agent. To deploy one, see Deploy a hosted agent.
• The azd Foundry extensions installed. For installation steps, see Install the azd Foundry extensions.
• An authenticated Azure Developer CLI session. Run azd auth login if needed.
• For session-specific logs, a session ID from an azd ai agent invoke response. To invoke an agent, see Invoke a hosted agent with the Azure Developer CLI.

View recent console logs

• Fetch recent console logs:

  azd ai agent monitor
  

  This command fetches recent console logs, including stdout and stderr, from the agent's last invoke session. If no session exists, it streams container logs. The command exits after fetching the available logs. Use --follow to stream continuously.

Stream logs in real time

• Stream logs continuously:

  azd ai agent monitor --follow
  

  Press Ctrl+C to stop. This is the most useful mode for debugging. Run it in one terminal while sending requests in another.

View system event logs

• Show container lifecycle events instead of console output:

  azd ai agent monitor --type system
  

  Use system event logs to diagnose container crashes, restart loops, and resource issues.

View session-specific logs

• Filter logs to a specific agent session:

  azd ai agent monitor --session-id <session-id>
  

• Combine with --follow for real-time streaming:

  azd ai agent monitor --session-id <session-id> --follow
  

  To find session IDs, check the output of azd ai agent invoke. It prints the session ID for each request.

Control log length

• Show the last 100 lines:

  azd ai agent monitor --tail 100
  

  The range is 1-300. The default is 50.

Monitor a specific agent

• In multi-service projects, pass the agent name:

  azd ai agent monitor my-agent
  

Recognize common log patterns

Pattern	Meaning	Action
Listening on 0.0.0.0:8088	Agent started successfully.	None needed.
AuthenticationError	The agent's Entra Agent Identity can't authenticate.	Check RBAC roles.
ModelNotFound	Model deployment name mismatch.	Verify deployment name in agent.yaml matches Foundry portal.
ResourceNotFound	Foundry endpoint mismatch.	Check FOUNDRY_PROJECT_ENDPOINT value.
Container restart events in system logs	Crash loop.	Check code for unhandled exceptions; consider increasing container resource limits in azure.yaml.
TimeoutError	Request took too long.	Check model responsiveness; increase timeout on invoke.

Follow a debugging workflow

A typical debugging session looks like this:

1. Stream logs in one terminal:

   azd ai agent monitor --follow
   

2. Send a request in another terminal:

   azd ai agent invoke "Test message"
   

3. Watch the logs for error patterns or unexpected behavior.

4. Check system events if the agent seems unresponsive:

   azd ai agent monitor --type system
   

For a comprehensive debugging workflow, see Debug a hosted agent.

Related content

• Debug a hosted agent for step-by-step diagnostic workflows.
• Test a hosted agent for validation strategies before production.
• Isolate hosted agent sessions per user for logs in isolated sessions.