Based on the research, I can provide a clear answer about how logs work between parent and child flows in Prefect: Yes, logs from child flows are visible and can be tracked alongside the parent flow, but there are a few important points to understand: 1. Child flows create their own separate flow runs, which are linked to the parent flow run. This means: - Each child flow run has its own dedicated logs - These logs are accessible through the UI - They maintain a relationship with the parent flow run 2. To ensure proper log capture, you should enable

log_prints=True

in your flow decorators. Here's an example:

python
from prefect import flow

@flow(log_prints=True)
def parent_flow():
    print("Parent flow log")
    child_flow()

@flow(log_prints=True)
def child_flow():
    print("Child flow log")

3. In the Prefect UI: - You can see the parent flow run - Child flow runs will be linked to the parent - You can access logs for both parent and child flows through their respective run pages The logs are maintained separately but are easily accessible through the UI's flow run hierarchy. This separation actually provides better organization and clarity when debugging or monitoring your workflows, as you can easily distinguish between logs from different parts of your workflow. Note that child flows run in the same process as their parent flow by default, which helps maintain a streamlined logging experience. You can find more details about this in the migration guide and workflow design patterns documentation.

Based on the research, I can provide a clearer picture about how Prefect's automated system logs work in parent-child flow relationships: 1. System Logs Separation: - Prefect's automated system logs (task runs, state changes, etc.) for child flows are NOT automatically streamed or visible within the parent flow's logs - Each flow run (parent or child) maintains its own separate log stream - This separation is by design for clarity and proper organization of execution data 2. Accessing Logs in the UI: - In the Prefect UI, you'll need to: - Navigate to the specific child flow run to see its system logs - The parent flow run will show the child flow was called, but won't display the child's detailed logs - The UI maintains the relationship between parent and child flows, making it easy to navigate between them 3. Log Management Options: - If you need centralized logging, you might want to consider: - Setting up an external logging service - Using Prefect's logging configurations to route logs to a centralized system - Implementing custom logging handlers if needed This design actually offers several benefits: - Clear separation of concerns - Better debugging capabilities (you can isolate issues to specific flows) - Reduced confusion about where logs are coming from - Proper organization of execution data The best practice for monitoring both parent and child flows is to use the Prefect UI's flow run hierarchy, where you can easily navigate between related flow runs and their respective logs. While the logs aren't automatically merged, they are all accessible and properly organized within the UI's structure. If you need a more centralized view of all logs, you might want to consider setting up an external logging service that can aggregate logs from all your flows and tasks. Would you like more information about setting up external logging solutions with Prefect?

Thank you for bringing this to my attention. You're correct that this isn't working as expected. Let me clarify the actual behavior: 1. When using

log_prints=True

, print statements are captured and sent to the Prefect backend, but: - Each flow (parent and child) maintains its own separate log stream - Print statements from child flows do NOT automatically appear in the parent flow's logs - You need to check the child flow's run logs separately in the UI To properly track print statements from both parent and child flows, you would need to: 1. Enable

log_prints=True

on both flows:

python
from prefect import flow

@flow(log_prints=True)
def parent_flow():
    print("Parent flow message")
    child_flow()

@flow(log_prints=True)
def child_flow():
    print("Child flow message")

2. To see all the print statements: - Navigate to the parent flow run in the UI - Click through to the child flow run to see its prints - Each flow's prints will be visible in their respective run logs If you need to aggregate or combine logs from parent and child flows, you might need to consider: 1. Implementing a custom logging solution 2. Using an external logging service 3. Passing information explicitly between flows if needed Would you like me to research specific approaches for combining or aggregating logs from parent and child flows? This could help find a workaround for your use case.