We Added One Bot's Name, and Three Automation Scripts Quietly Stopped
Last week we formally registered a new bot named Codi into the dream team. Adding a name to the dashboard and one line describing its role felt like a five minute task.
The problem surfaced the next morning. Three automations, card news publishing, the Slack morning greeting, and the inventory report, had quietly stopped at the same time slot. No error logs were left behind, so at first we thought the server had died.
It took over an hour to find the cause. The bot name list did not live in one place. It was scattered across three files, a JSON file for the dashboard, a hardcoded list buried inside the automation scripts, and a mail sending config file. We had only added the new name to the dashboard, leaving the other two untouched. The scripts read Codi as an unregistered bot and quietly excluded it from their execution conditions.
Mannungi, our orchestrator, was digging through the coordination logs when it noticed all three scripts failed at the same moment for the same reason. One missing name had managed to bring down three unrelated looking systems at once.
We patched all three spots by hand to restore service, but we did not stop there. We consolidated the bot name list into a single file and rewired every script to read from that one source. Now adding a new bot only requires editing one place.
That day we learned that a small name can freeze three systems at once. Keep the same information scattered across multiple places, and eventually it will drift out of sync.