RDD troubleshooting
Check the saved policy, connection diagnostics and operation outcome before repeating an action.Connection checks
Check whether the bridge is enabled, the Raindrop bot is available, the guild and selected resources still exist, and the bot has effective permissions. Resource IDs after a diagnostic help identify the failing selection. Restore the intended permission or selection, then repeat a small positive check. Do not use Administrator to conceal a missing permission.
Settings did not apply
A 409 revision conflict means another save or state change won. Refresh, compare the draft and confirm again. A stale response must not replace a newer policy. Confirm that sibling servers share the intended primary and guild settings. Do not change the binding to a different guild to work around an error.
Local browser URLs
For local or self-hosted testing, set web.baseUrl in plugins/RDD/config.yml and ask the API operator to set app.frontend.url (APP_FRONTEND_URL) to the same WEB origin. RCore central.baseUrl is the separate API address. HTTPS origins and HTTP on localhost, 127.0.0.1 or [::1] are accepted; credentials, paths, queries and fragments are rejected. Browser localhost refers to the browser’s own computer. Ordinary hosted users keep the hosted defaults.
# plugins/RDD/config.yml
web:
baseUrl: http://localhost:3000
# API environment
# APP_FRONTEND_URL=http://localhost:3000Rejected startup or reload
A fresh install creates the default configuration file. An existing file must be readable, valid YAML with a string web.baseUrl. Invalid startup configuration leaves RDD disabled. A rejected /rdd reload, including a missing file, keeps the last valid browser links. Fix the file and restart a disabled plugin/server; use /rdd reload for a running plugin. Changing the API environment setting requires an API restart. Invalid API origins return an unavailable reply.
Queued, expired or uncertain operations
Check the exact server and receipt before repeating a command or reward. UNKNOWN means execution is uncertain, not proof that nothing happened. Never replay it automatically. Ordinary events expire after a minute; ban/unban events can survive outages for up to seven days. Status metadata may wait for its configured interval. Redact keys, tokens, cookies, OAuth codes and webhook URLs when asking for support.
Disable and clean up safely
Stop new console, reward and moderation triggers, then disable the intended policies or bridge through the editors. Keep the bot, database ownership records and required permissions available while cleanup reconciles owned roles, nicknames, voice channels, overwrites and mutes. Preserve manual grants and later human nickname changes. Do not delete resources by an RDD name prefix or remove the bot before cleanup finishes. Restore access and contact support if cleanup remains pending.