Question

Hi, i'd like to know how does ngx_http_upstream_dynamic module interact with nginx failure handling mechanisms.

When defining multiple servers nginx can know when one is unavailable and fallback to the next server. In our case we have an upstream composed of a single server which is a ELB and resolves to multilpes IPs, if we configure failure thresholds for this single servers will tengine fail to whole upstrem when a single backend IP fails? or will tengine consider them as separate servers and disable only the failing IP?