Since switching from the LogDNA addon to log streams a few days ago, I’ve experience a half dozen or so failures (and thanks for the automated emails!). It seems like a misconfiguration would have led to a hard failure, rather than the current intermittent one. Additionally, I would find it surprising – though not unbelievable – if this were an issue with LogDNA, given that their core business is high-uptime logging.
Is there any chance this is a render-side issue? If not, any suggestions for how to fix it?