Currently the way HA dispatch is set up, it will mark a url failover only when a retry is attempted. This causes two issues
- When a request fails (say request-1) on the last retry attempt that host (say host-a) is not marked as failover so when a new request (request-2) is made this host (host-a) will be chosen.
- There could be a use case where the services do not want automatic failover but would like to have new request reach a new host (e.g. by marking retry=0)