No server to go down
Dispatch is a Cloudflare Worker, not a process on a machine. There is no host to reboot, patch or run out of memory. Each request runs in a fresh, isolated sandbox, and a crash in one never touches the next.
Resources
Numbers from the running bot, and the reasons it stays up.
Cloudflare Workers
—
Discord servers with Dispatch installed.
—
Across those servers. Someone in two servers counts twice.
—
Answering from the nearest edge.
—
Cloudflare D1 round trip.
—
Time to read the server list.
Checking…
Dispatch is a Cloudflare Worker, not a process on a machine. There is no host to reboot, patch or run out of memory. Each request runs in a fresh, isolated sandbox, and a crash in one never touches the next.
Most bots hold a gateway connection open all day and have to reconnect when it drops. Dispatch has none. Discord sends each command to an HTTPS endpoint, so there is nothing to reconnect and nothing to miss while we deploy.
Cloudflare runs the Worker in hundreds of cities and answers each request from the nearest one. A traffic spike is spread across the network instead of piling onto one box.
Discord gives a bot three seconds to respond. Anything that touches the database or another API replies right away with “thinking…” and finishes in the background, so a slow query never turns into a failed command.
Reports live in Cloudflare D1, which keeps storage durable and backed up, with point-in-time restore. Attachments are rehosted in R2, so a report still has its screenshots after the original Discord message is deleted.
Every change goes through typechecking and an end-to-end test suite that runs real signed interactions through the Worker before it merges. Deploys swap in a new version without a restart.
This page shows what is happening right now, not a promise. If something is wrong, open an issue.
Add Dispatch to your server and connect a repository in a couple of minutes.
Add to Discord