SLES for SAP workloads

Assumption: I am using your question is for SAP NetWeaver AS ABAP based application on HANA Database (eg ECC, S/4)

1. SLES for DEV and QA Instances deployed in Single AZ Answer: Yes, this is how most Customers. Make sure your SLA is met with this design. See https://aws.amazon.com/compute/sla/
2. SLES for SAP Prod application servers deployed in single AZ and replication to other AZ/Region using AWS elastic DR Answer: You can do both ways. It depends upon your RTO & RPO plus the use case. Are you looking for a HA solution or a DR solution? Elastic DR service is primarily for cross-region DR use cases but technically it can be used in the same region across two AZs. Since the latency is very low between the two AZs in the same region, a better approach will be to distribute your application across AZs and also setup an ERS and HANA secondary node in the second AZ. You can automate the failover using SLES pacemaker cluster. But again, there isn't information provided here to provide a recommendation.
3. SLES for HANA DB(in AZ 1 and AZ2) . AZ 2 HANA instance configured for lean DR replicating using HSR. Answer: Yes, you can do this if your HA RTO requirement is met. The secondary instance should still have enough memory to accommodate your row store table plus another 50GB for runtime. Also, adjust the HANA parameters to turn off table preloads.

Refer to the following guide for more details: https://docs.aws.amazon.com/sap/latest/general/general.pdf#architecture-guidance-of-sap-on-aws