captureException can cause thread blocking on server

[justin--]

Hey everyone. Trying to figure out whats going on. In some situations where we send an exception back to back, we notice that our nodejs server will slow down quite a bit. Something that loads within a second will take 2+ seconds, and if something generated a lot of errors at once it could block the server for 10s+.

Has anyone experienced this? I've done this with the following test middleware. (this is an extreme example of doing it on every request, but it can be reproduced in isolation on a single request awhile trying to do other requests).

import { captureException, captureMessage, init } from '@sentry/node';
import type { NextFunction, Request, Response } from 'express';

init({
  dsn: 'https://[email protected]/1',
  debug: true,
  tracesSampleRate: 1.0,
  environment: 'test',
  enableTracing: false,
  autoSessionTracking: false,
  sendClientReports: false,
  attachStacktrace: false,
});


export function exampleMiddleware(req: Request, res: Response, next: NextFunction): void {


  const start = new Date();
  const error = new Error('Test Exception');
  captureException(error);
  const end = new Date();
  console.info('Timing:', end.getTime() - start.getTime());

  return next();
}

I've tried changing different settings, and even pointed to the public sentry server to rule out being my company's sentry server. I'm a loss because this can easily cause latency issues for us, making me want to stop using it for server error tracking.

[AbhiPrasad]

Hey @justin--, if you drop the event with beforeSend option in Sentry.init do you still see this behaviour?