Metrics & monitoring
Read CPU, memory, disk, network and uptime charts for an instance across every time range the panel offers.
Time ranges
The Metrics tab shows the same data across five ranges:
| Range | Use it for |
|---|---|
| Live (1H) | Watching the effect of a change right now |
| 24 Hours | Daily patterns and recent incidents |
| 48 Hours | Comparing today with yesterday |
| 7 Days | Weekly load patterns |
| 30 Days | Capacity planning and long-term trends |
Live is near-real-time and refreshes as it runs. Longer ranges are aggregated, so short
spikes can be flattened: use the live view when you need the highest resolution (refreshed
every 10 seconds).
Metric groups
- Uptime: whether the instance was reachable, and gaps in collection.
- CPU Usage: processor load over time; sustained high values mean the plan is too
small for the workload. - Memory usage: RAM used by your server; watch for steady growth that never drops,
which usually means a leak. - Disk space: used capacity; plan for headroom before it reaches 100%.
- Disk I/O: read and write activity; large values alongside slow applications point
at the disk rather than the CPU. - Network (public and private): throughput on each interface.
- Network packets (public and private): packet rates, useful for spotting floods or
unusual traffic patterns.
Limits of panel metrics
- VoxiHost collects metrics at the platform level; nothing needs to be installed inside your
instance. - A stopped instance keeps its history (availability shows 0), but live charts stop updating.
An instance that is still being created has no data yet. - The panel measures the instance, not your application. It cannot show per-process usage,
request latency, query times or business metrics. - There is no alerting. If you need notifications, run a monitoring service on the instance
or outside it; /guides/ covers options such as Netdata and Uptime Kuma. - History is kept for 30 days; save anything older yourself. The panel has no metrics
export.
For firewall and network-level events, use the Firewall tab and the
Firewall rules documentation. For host-level incidents,
open a support ticket with the time window and the chart range you were looking at.