• Review the documentation: https://docs.sentry.io/
• Search for existing issues: https://github.com/getsentry/sentry-javascript/issues
• Use the latest release: https://github.com/getsentry/sentry-javascript/releases
• Provide a link to the affected event from your Sentry account

Package + Version

• @sentry/browser
• @sentry/node
• raven-js
• raven-node (raven for node)
• other:

Version:

5.15.5

Description

For the longest time I've been passing an onFatalError option where I do something along the lines of what Raven used to do:

https://github.com/getsentry/raven-node/blob/56a588e3deab4494c6f41e14ed906f3d40511781/lib/client.js#L86-L89

    this.onFatalError = this.defaultOnFatalError = function (err, sendErr, eventId) {
      console.error(err && err.stack ? err.stack : err);
      process.exit(1);
    };

(For context, the only reason I did this was just because I wanted to prefix the logged error message with "fatal error", to differentiate them from non-fatal errors we choose to log.)

However, just today I noticed that uncaught exceptions were not being reported to Sentry. After some debugging it seemed to be because of my custom onFatalError. In the latest version of @sentry/node, the default onFatalError only kills the process when all events have finished sending:

By passing a custom onFatalError I was unintentionally overriding this behaviour, thus the event never reached Sentry because the process was killed before the event was sent.

For this reason, I think the docs for onFatalError should carry a big warning—that by customising this functionality, you lose this important facility. WDYT?