DNS issue with Rancher/Rancher controller node

Attempts to create pipelines attached to GitHub or Bitbucket (cloud hosted) are consistently failing due to UDP failures (running v2.3.3)

Post https://bitbucket.org/site/oauth2/access_token: dial tcp: lookup bitbucket.org on read udp> i/o timeout

I managed to successfully create one pipeline after performing a rolling restart of my Rancher worker nodes (and controller), though I ended up creating it in the wrong project and attempting to create in the correct project went back to the error above. After rebooting the controller node the problems cleared. A day later I created a new GIT repo and did a ‘Refresh Repositories’, met with the same UDP timeout error. Rebooting the controller node cleared the error and the refresh completed nearly instantly, with the new repo properly displayed.

I only see one container (rancher/rancher) running on the controller node, so I am not sure what process is hanging (or crashing/stopping) preventing DNS lookups from properly executing after a given period of time.