In my existing monolith application, I generate files that need to be sent as email attachments. I've started moving to a microservices architecture, where an event is raised via RabbitMQ to trigger email sending in a different microservice.
The challenge is: I need to send the file generated by Monolith as an attachment in the email from the new microservice.
I know that using an S3 bucket (or similar object storage) would be a good approach in a cloud environment—upload the file, pass the reference via the event, and download it in the email service. However, I also need to support an on-premises deployment, where I can't rely on services like S3 or introduce additional infrastructure that clients must maintain.
What is the best way to handle this scenario in both cloud and on-prem environments?
What would you recommend for this kind of hybrid setup?
You're right - in the cloud, the solution is to upload the file to object storage
(like S3, MinIO, or Azure Blob Storage), then send a reference (URL or ID) to the email service via your RabbitMQ event.
But for on-premises setups where external infrastructure like S3 isn't available or desirable, you can use a hybrid approach:
In the cloud:
fileId in the RabbitMQ event, and let the email service download the file using that ID.On-prem:
/var/app/files/report.pdf) in the event, and let the email service read the file directly from that location.The email service is subscribed to the RabbitMQ queue, receives the event, and loads the file either from object storage (in the cloud) or from the local filesystem (on-prem) before sending it as an email attachment.