When running Sentry with NextJS, what is the best way to uncover minified server errors?

[jonschmidt]

When an error is thrown in a server component in nextjs, we get the following

Error
An error occurred in the Server Components render. The specific message is omitted in production builds to avoid leaking sensitive details. A digest property is included on this error instance which may provide additional details about the nature of the error.

It includes a "digest" that we can use to find the real error.

Right now, we are able to take this error, ssh into our instance, and grep our logs for the digest, and then we find something like

/var/log/web.stdout.log-5123-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  code: 'OAUTH_RESPONSE_BODY_ERROR',
/var/log/web.stdout.log-5124-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  error: 'invalid_grant',
/var/log/web.stdout.log-5125-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  status: 403,
/var/log/web.stdout.log-5126-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  error_description: 'Unknown or invalid refresh token.',
/var/log/web.stdout.log-5127-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  [cause]: [Object]
/var/log/web.stdout.log-5128-Apr 14 05:42:27 ip-172-31-10-154 web[204747]: }
/var/log/web.stdout.log-5129-Apr 14 05:42:27 ip-172-31-10-154 web[204747]: ⨯ Error: Server response error: 401
/var/log/web.stdout.log-5130-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:    at <unknown> (.next/server/app/[locale]/users/[slug]/route.js:1:4010)
/var/log/web.stdout.log-5131-Apr 14 05:42:27 ip-172-31-10-154 web[204747]:    at <unknown> (.next/server/app/[locale]/users/[slug]/route.js:1:4133)
/var/log/web.stdout.log:5133:Apr 14 05:42:27 ip-172-31-10-154 web[204747]:  digest: '271938285'

So while this tells us the actual error that was thrown, the code is minified, and we cannot accurately figure out where it is being thrown from.

One thing I think I saw someone mention is that if you have tracing enabled, you can see the actual error further up the trace. So is it true that the only way to get the actual source of these server component errors with the proper source code is to pay for tracing?

Any other tips here?

Thanks

[lforst]

If you enable tracing in the SDK, Sentry should show you the related error in the backend via the trace.