Error Handling

Customizing the Error Message

Bottender makes it easy to display custom error message for various runtime errors. To customize the error message sending to the user, you may create an _error.js file in your project root:

// _error.js

module.exports = async function HandleError(context, props) {
  console.error(props.error);
  // or you can choose not to reply any error messages
  await context.sendText(
    'There are some unexpected errors happened. Please try again later, sorry for the inconvenience.'
  );
  if (process.env.NODE_ENV === 'production') {
    // send your error to the error tracker, for example: Sentry
  }
  if (process.env.NODE_ENV === 'development') {
    await context.sendText(props.error.stack);
  }
};

The example code above will do following things for you while an error occurred:

1. Send There are some unexpected errors happened. Please try again later, sorry for the inconvenience. text message to the end user.
2. Log error to the console.
3. Log error into error tracker.
4. Send error.stack as text message to the user (only in the development)

Note: In LINE channel, errors should be handled in a different way because reply API can only be called once. You can read more about it in the Error Handling in LINE documentation.

Sending Errors to Sentry

Sentry is an error tracking and monitoring tool that aggregates errors across your stack in real time.

To integrate with it, first, you need to download the sdk from registry to use it:

# Using npm
$ npm install @sentry/node

# Using yarn
$ yarn add @sentry/node

Then, add those few lines of code in your _error.js to send your runtime errors to Sentry. Make sure that you fill in you DSN (Data Source Name) to configure the Sentry instance:

// _error.js
const Sentry = require('@sentry/node');

Sentry.init({
  dsn: 'YOUR_SENTRY_DSN',
});

module.exports = async function HandleError(context, props) {
  if (process.env.NODE_ENV === 'production') {
    Sentry.captureException(props.error);
  }
};

When errors happen in production, those errors will be sent to Sentry automatically.