Events

Everything a workflow does starts with an event. Understanding your options for what starts a run — and how workflows can signal each other — unlocks most of the interesting builds.

Three kinds of starting points

  • Discord happenings. A member joins, a message arrives, a reaction is added, a button is clicked. Pick the trigger, and the event's details — who, where, what — come out of the trigger block as outputs ready to wire.
  • Schedules. Run every morning, every hour, or once at a set time. Good for digests, cleanups, reminders, and anything not driven by a person.
  • Events from other workflows. Workflows can announce things, and other workflows can listen. This is the glue for multi-workflow builds.

A workflow can have several triggers — a ticket workflow might react to a button and a slash command and a nightly cleanup schedule, sharing the same functions underneath.

What your server needs

Some Discord triggers only fire if the server grants the right access — Discord gates certain event streams and actions behind permissions and intents. You don't need to memorize any of it: the editor reads your graph and shows one combined list of what the workflow needs before you publish. If a trigger silently never fires on a server, this checklist is the first place to look.

Announcing your own events

Give your workflow a voice by declaring an event of its own — name it, give it a payload (the values that travel with it), and emit it from the graph like any other action. Mark it public and other workflows can trigger on it.

This is the pattern for keeping big builds sane: instead of one giant workflow that does everything, a core workflow announces "ticket opened" and "ticket closed", and any number of smaller workflows react — a logger, a stats counter, a satisfaction survey — each publishable and testable on its own.

Delivery you can trust

An emitted event isn't a shout into the void. Each listener gets its own delivery, failures retry automatically, and a delivery that keeps failing is parked — with its error — where you can see it in Monitoring, instead of vanishing. One broken listener never stops the others from getting their copy.

Changing an event later

Once other workflows listen to your event, its payload is a promise. Adding a new field is safe — existing listeners ignore it. Removing or reshaping fields breaks listeners, so Juno makes payload changes explicit: you declare whether a change is additive or breaking, and dependents see that signal when they upgrade rather than finding out the hard way. When in doubt, add a new event alongside the old one.