Error forwarding

The plugin catches unhandled errors (uncaught exceptions and unhandled promise rejections) and forwards them to a Discord channel. Each error appears as a rich embed with the full stack trace and two action buttons: Dismiss and Restart.

This only runs in development mode (robo dev). In production, the entire plugin is stripped from the build.

Setup

Point the plugin to a Discord text channel. You can do this with an environment variable:

DISCORD_DEBUG_CHANNEL_ID=1234567890123456789

Or through the plugin config:

export default {
	debugChannelId: '1234567890123456789'
}

export default {
	debugChannelId: '1234567890123456789'
}

The config option takes priority over the environment variable.

How it works

The plugin registers an error hook at src/robo/error.ts. When the Robo framework catches an unhandled error, this hook fires and calls sendDebugError() which:

1. Checks that the Robo is running in dev mode
2. Reads the debug channel ID from config or env
3. Gets the Discord client from @robojs/discordjs
4. Sends a red embed to the channel with the error stack trace
5. Attaches Dismiss and Restart action buttons

If any step fails (no channel configured, client not ready, channel not found), the function returns false silently. Errors during forwarding don't crash the process.

Interactive buttons

Each error embed includes two buttons:

The restart behavior depends on how the Robo is running:

• robo dev mode — Sends a { type: 'restart' } IPC message to the parent process for a graceful restart
• Standalone — Calls process.exit(0) and relies on a process manager to restart

Role pinging

Notify your team when errors occur by configuring a role to ping:

export default {
	debugChannelId: '1234567890123456789',
	errorPingRoleId: '987654321098765432'
}

export default {
	debugChannelId: '1234567890123456789',
	errorPingRoleId: '987654321098765432'
}

When an error is forwarded, the message includes a role mention (<@&roleId>) that pings all users with that role.

Programmatic usage

You can also send errors to the debug channel from your own code:

import { sendDebugError } from '@robojs/dev'

try {
	await riskyOperation()
} catch (error) {
	const sent = await sendDebugError(error)
	if (!sent) {
		console.error('Failed to forward error:', error)
	}
}

import { sendDebugError } from '@robojs/dev'

try {
	await riskyOperation()
} catch (error) {
	const sent = await sendDebugError(error)
	if (!sent) {
		console.error('Failed to forward error:', error)
	}
}

For replying to interactions with error embeds, use printErrorResponse:

import { printErrorResponse } from '@robojs/dev'

export default async function (interaction) {
	try {
		await someCommand(interaction)
	} catch (error) {
		await printErrorResponse(error, interaction)
	}
}

import { printErrorResponse } from '@robojs/dev'

export default async function (interaction) {
	try {
		await someCommand(interaction)
	} catch (error) {
		await printErrorResponse(error, interaction)
	}
}

This handles all interaction states — deferred, already replied, or fresh — and sends an ephemeral red embed with the error message.

See the API reference for full function signatures.

Stack trace truncation

Discord embeds have a 4000-character limit for descriptions. Stack traces longer than this are truncated automatically.

Next steps