I’m running into a sporadic issue with a set of mongo databases that are sitting behind an AWS Network Load Balancer.
These are treated as read only databases, each one having an identical copy of the same set of data, while the network load balancer itself has been configured for client affinity (per the production deployment guidelines).
However, every once in a while (roughly twice per hour while the connected service is under load - which is also sitting behind its own Application Load Balancer), an exception is logged regarding being unable to find a cursor when performing one of the queries we service.
I am wondering if anyone has had any experience setting a system like this before, and if this is a problem you may had run into. Appreciate any help you can provide, and thanks in advance.