There were some brief periods starting at 09:19 UTC where an upstream provider had issues causing connection problems to our backend instances. This resulted in Orb clients automatically reconnecting until they got a stable connection. These spikes were finally resolved at 12:42 UTC.