Lost redirect via HTTP503, reverted back to a HTTP302
During the 2018-07-31 failover test on staging:
After we failed over to GCP, we lost the HTTP503 that was being served on the gstg.gitlab.com
- instead it started serving an HTTP302
need to check LBs in gprd/gstg to make sure config is correct (azure was updated to do 503, presumably gcp pre-existing config has not been updated)