At 4:30 PM PST our customers started to notice extreme slowness with all parts of our site and around 4:50 PM PST we declared an incident.
During investigations into the root cause we determined that GitHub was suffering from extreme latency for some of their API's. This latency was having an impact on CircleCI at all levels which translated into site slowness and/or build failures.
GitHub resolved their issue around 5:50 PM PST which then also returned out site back to normal.