Very high number of Internal errors

[raimille1]

We deployed to production Monday night and we're seeing a concerning amount of errors (error type Internal) that we're not even able to trace on baselime.io.

We're blind as to how to fix them since logs do not show anything more than just that there are errors.

Very concerned this will affect our bill as well, given that the clients are reconnecting and we're spending GB-sec and storage read/write operations.

After just a few days of operation, we've observed 1.3Million errors out of 2.2Million requests.

cc: @threepointone