It's a matter of scale. When you had a few in-house servers, somebody might manage their events and exceptions. Maybe 10 a day.

When it went to the cloud, complexity rose and your visibility suffered (you had little insight into what hardware events meant on a virtual server). Maybe 100 a day.

When servers grew to hundreds, you had to use tools to manage the logs, summarize, prioritize, spot commonalities.

Thousands of servers meant now you use tools to manage the tools? Too many details even in summaries; deploying fixes takes scripts or scripts of scripts.

I sometimes think, AI isn't going to ruin our civilization. AI is our only hope for surviving the geometric growth in our civilization's complexity.