A ¥1.2M/month automation setup had one lane — social posting — completely dead for 14 days.
And the whole time, the system reported "healthy" every morning. No errors anywhere. Script exit codes were 0. Every watchdog was green. Yet Instagram hadn't published a single post since July 29, and TikTok since July 28.
I only found out because I was digging through logs for something else.
Why This Approach Works
Once you've run automation for a while, you develop a specific dread about this class of failure. Systems that crash loudly are easy to fix. The problem is systems that die quietly.
