[FIXED] Internal Server Error (possibly linked to Google SSO)

[saas_dev]

A few hours ago (around 9am US pacific) we started seeing internal server errors popping up when users attempted to log in to one of our Anvil apps through Google SSO. Here’s an example of the error (the specific error ID changes every time someone attempts to log in) and the code that it points to:

The app has not been edited in a few weeks, and the errors only began this morning. We have one other Anvil app that uses Google SSO and is working fine now (although experienced a separate and resolved issue with anvil.tables in the same time window).

From the context, it doesn’t look like something we can trouble shoot on our end. Is this a known issue? Any tips on how to solve it?

[meredydd]

Hi there,

Once you’ve hit an “Internal server error” like that, further investigation needs to happen on our end - that hex code leads us to internal diagnostic logs – which, in this case, were enough for us to narrow down the issue and ship a fix! Apologies for the inconvenience – you should now be back on the road. Let us know if not.

(If you’re interested, it turns out that the large number of columns in your tables hit an internal limit in Postgres! We’ve implemented a workaround, and will be expanding our test suite to make sure we test the “very large numbers of database columns in a single fetch” case )