Grafana Log Dashboards — From Fleet Overview to Per-Host Detail

#grafana #loki #linux #sysadmin

Email-based monitoring has a failure mode: the inbox fills up, mail rules get created, and eventually something critical gets missed.

Two Grafana dashboards break that pattern.
The three-screen flow:

→ Network Overview: error count column, red when problems exist, click to investigate → Log Summary: errors, warnings, failed logins, invalid users, sudo events, total log lines → Host Logs: log volume time series + three pre-filtered log panels
What's new on Host Logs:

✓ All Logs — full unfiltered stream for context ✓ Errors & Warnings — pre-filtered, first stop when stat panels are red ✓ Auth & Security — raw auth log with user, IP, and timestamp ✓ Annotations on the time series for marking events during investigation
The design principle:

Each screen narrows the scope. Two clicks from signal to log lines.