At 1:12pm EST we deployed an infrastructure upgrade that blocked requests in production. Engineers were alerted and we immediately rolled back. Traffic was restored at 1:17pm.
