Atlassian has identified a faulty server as the cause of these errors. The faulty hardware has been removed from their load balancer and everything should be back to normal.
Since Monday, April 1 at 12:30pm UTC, we are seeing a much-elevated number of connection errors to the Jira REST API, resulting in post-function failures. We are currently investigating this with Atlassian.