Plays/Automation & Routines

PlayOther tools

Answer the messages a support bot missed during a server outage, then switch the catch-up off

Freelance automation developer Dmitry Prozorovsky runs an n8n-based AI support bot for a motorcycle-gear shop's VK community.

Reported by Dmitry Prozorovsky (via UptimeRobot)on x.com

A prompt to try

Design an n8n recovery workflow for my support bot:
1. Trigger: poll [MONITOR] API every [N] minutes for my server's recent up/down history.
2. Start the backfill only after the server has been up for [N] minutes in a row.
3. Pull conversations from [CHANNEL] for the outage window and compare against the bot's reply log; keep only messages with no reply from the bot or a human.
4. Answer each one, fetching current order status and prices at reply time.
5. When the list is empty, deactivate the backfill workflow.
List the nodes, and how to avoid touching conversations a human is already handling.

Fill in anything in [brackets] before you send it. Anything that spends money, sends a message or signs something should stay behind your own approval.

How it went, step by step

  1. 01Watch the server from outside it, so the check still works when the server is down.
  2. 02After an outage, wait until the monitor history shows the server is stable again.
  3. 03Compare conversations with the bot's own records and reply only to messages that never got an answer.
  4. 04Look up live order status and prices at reply time, then switch the workflow back off.
Read the original source Discuss in the forum