When the lead developer notices that MTA child processes are hanging, it could indicate that the maximum number of MTA child processes is too low in relation to the allowed maximum number of pending Delivery Parts per MTA child. When the number of pending delivery parts exceeds the maximum number of child processes, the child processes will not be able to process all the pending parts, leading to a delay in delivery and the child processes hanging.

This issue occurs when the wfserver process is not set to automatically start when the server is restarted. This means that the workflows will not run unless the wfserver process is manually started. The Start/Restart button will be disabled and the process will not run until it is manually started.

The architect should recommend configuring the Adobe Analytics Genesis connector, which is a tool that allows for integrating Adobe Analytics data with other marketing systems, such as email and SMS campaigns. By leveraging the remarketing workflows in the Adobe Analytics Genesis connector, the retail company can target customers who have made product complaints and execute a windback campaign to them within 1 hour of their web session. This approach allows for real-time triggering of the campaign based on customer behavior, which is essential for executing a campaign within 1 hour of a customer's web session.

Validate if the delivery can be paused and the MTA process restarted to clear the personalization cache and reuse the delivery. This approach will allow the Architect to pause the delivery and restart the MTA process, which will clear the personalization cache and allow the delivery to be reused. Additionally, the Architect should validate if the delivery is complex with many JavaScript personalization and whether the email size is more than 60 kb, if all images can be uploaded into public resources to fetch faster, if personalization blocks can be created in the delivery, and if the number of recipient per delivery can be reduced to send multiple deliveries at the same time to spread the load on personalization.