Description
We lack documentation on high availability and disaster recovery.
This is not intended to write down how to set up high availability in Windows or Docker/Kubernetes, but rather the implications on messaging when setting up HA and disaster recovery.
One example could be what a customer did, set up high availability with two regions where not just the business data was replicated, but also messaging data inside SQL Server transport. This meant that on one region, messages could be processed, but the other region would pick up the same message at the same time. If Outbox is also replicated and both regions store Outbox info in their own region, it could happen that this data was created on both regions without them being aware that something happened on the other side. Replicating this data could theoretically cause even more issues.
More information can be found in Architecture guidance impact assessment issue.