Observability tools tell you what went wrong. They rarely help you fix it.
During an incident, Engineers bounce between Kibana, Slack, and the terminal/IDE often. Every tool switch, the assistant stays behind in the previous surface, and the investigation restarts from scratch🧵