We just resolved the issue!
We implemented temporary fix while we work out a robust solution and currently monitoring the logs.
We are currently investigating this incident.