Discord API rate limits for bots
Discord rate limits protect the platform. Bots that ignore them look randomly broken. This guide explains practical patterns that keep community bots stable.
Where limits show up
REST calls (messages, moderation actions, command registration) and some gateway behaviors are limited. Libraries usually surface HTTP 429 responses and retry-after hints. Ignoring them and retrying immediately makes the problem worse.
Batch and queue work
Mass role changes, channel wipes, or bulk messaging should be queued with concurrency limits. Prefer Discord bulk delete where appropriate instead of tight loops of single deletes.
Caching
Cache immutable or slow-changing data in memory with TTLs so you do not refetch on every command. Invalidate on relevant events instead of polling aggressively.
User-facing errors
When a limit is hit, tell moderators the bot is cooling down instead of failing silently. Log the route and guild for later tuning.
Hosting angle
Rate limits are not solved by more RAM. They are API design issues. Free and paid hosts both need well-behaved clients—see uptime habits.