Skip to main content

rugmunchbot Deployment 🟢

Production checklist​

  • Bot token in secret manager (never in git)
  • Single instance per bot token (Telegram allows one active poller/webhook)
  • Redis for job queue persistence across restarts
  • Alert backlog monitored — a stuck scheduler silently kills the product
  • Error reporting wired (failed scans should surface, not vanish)

Scaling​

The bottleneck is Telegram rate limits, not compute. For multi-tenant setups: one bot process per bot token; share the RMI backend horizontally.