At a glance
Instance Uptime is the elapsed time since the MariaDB server process last started, shown in real time. A long, steadily climbing uptime is the boring-is-good signal a DBA wants. A sudden reset to near zero means the server (or the container hosting it) restarted, planned or otherwise, and is your first clue when chasing a connection storm, a cleared buffer pool, or a lost in-memory counter set.
| What it tracks | Seconds since mysqld / mariadbd last started on the selected instance, rendered as days, hours and minutes. |
| Data source | The Uptime global status variable (SHOW GLOBAL STATUS LIKE 'Uptime'). |
| Time window | RT (real-time, refreshed on the standard live poll). |
| Alert trigger | None. An unexpected reset is best caught via MariaDB Health Score and the connection/error cards rather than a raw threshold. |
| Roles | owner, engineering, operations |
What it tracks
The value comes straight from theUptime status counter, which MariaDB increments every second the server process is alive and resets to zero on every restart. Because so many other Vortex IQ MariaDB cards read from Performance Schema and SHOW GLOBAL STATUS counters that also reset at startup, Instance Uptime acts as the reference clock for the whole connector: if you see Queries per Second (live), Aborted Connects (24h) or buffer pool hit rate behaving oddly, check uptime first. A counter that looks abnormally low or a ratio that looks suspiciously perfect is often just a node that restarted minutes ago and has not yet accumulated a representative sample. On a Galera cluster a single node restarting is normal during rolling maintenance; all nodes restarting together points at an orchestration or host-level event.
Reconciling against the source
Confirm directly withSHOW GLOBAL STATUS LIKE 'Uptime'; (value is in seconds) or SHOW STATUS LIKE 'Uptime_since_flush_status'; if your counters were reset with FLUSH STATUS. On managed platforms cross-check against the host metrics in AWS RDS / Aurora, SkySQL or your container orchestrator’s pod-restart count; a process uptime that resets while the host stays up indicates a mariadbd crash or in-place restart rather than a full reboot.