A routine update in our deployment layer resulted in network connectivity issues between nodes. This resulted in services that were functioning normally to appear down to load balancing and proxy services. This ultimately resulted in the api being inaccessible externally.
Upon identifying the cause, we worked to re-establish connectivity internally, restoring load balancing and proxying services.
New maintenance processes have been put in place for the nodes in question. Preventive maintenance measures will take place on non-production nodes that will allow for testing and verification of the updates before they can impact production services. In production, these updates will then be applied during scheduled maintenance so that the impacts can be closely monitored.