Infrastructure Monitoring
General
Versio.io enables the collection and storage of metrics for analysing the dynamic behaviour of any entities.
- A Versio.io instance is required to store metrics (e.g. host ABC, sensor 123).
- Metrics can also be assigned to workspaces via their associated instances. This allows metric analyses to be used based on the workspace filter.
- All metrics are stored rounded to two decimal places.
- The following aggregated statistical metrics are available for each metric:
- Min
- Max
- Average
- Median
- Percentiles 99%, 95% and 90%
Sampling rate
The sampling rate for metrics depends on the metric source:
| Metric source | Sampling rate |
|---|---|
| Host | 1 minute |
| Network device | Customer-specific, min. 1 minute |
| Corroventa | Customer-specific, recommended 10 minutes |
| Printer | 15 minutes |
Downsampling & historisation
For space and performance reasons, the metrics are compressed over time or deleted depending on the Versio.io history. The maximum retention period for metrics is 365 days.
| Period | Downsampling | Number of data points per metric |
|---|---|---|
| 0-24 hours (1d) | - | 1.440 |
| 1-14 days (14d) | 2 minutes | 9.360 + 1.440 = 10,080 |
| 15-90 days (3m) | 10 minutes | 10.080 + 10.656 = 20.736 |
| 91-365 days (1y) | 60 minutes | 20.736 + 6.600 = 27.336 |
Custom metrics via API
Custom metrics (configuration) and metric values can be created via Versio.io Rest API and assigned to an instance. Metric values for an instance must always be more recent than the last stored metric value and may not be more than 2 hours old.
Supported metrics
The following standard metrics are available:
| ID | Name | Unit | Description |
|---|---|---|---|
builtin:host.cpu.usage |
% | Overall CPU usage as a percentage of total available capacity, including user and system time. | |
builtin:host.cpu.idle |
% | Percentage of time the CPU is idle and not executing any tasks (excluding I/O wait). | |
builtin:host.cpu.other |
% | CPU time not categorized as user, system, idle, or iowait—may include interrupt handling or guest time. | |
builtin:host.cpu.steal |
% | Time the virtual CPU waits for real CPU while the hypervisor is servicing other virtual machines. | |
builtin:host.cpu.iowait |
% | Percentage of time the CPU is idle while waiting for I/O operations (e.g., disk or network) to complete. | |
builtin:host.cpu.user |
% | Time spent on user processes (non-kernel code), including applications and services. | |
builtin:host.cpu.system |
% | Time spent executing system (kernel) processes and handling system calls. | |
builtin:host.cpu.interrupt |
% | Time spent handling hardware interrupts. | |
builtin:host.cpu.load |
number | ||
builtin:host.cpu.load5 |
number | ||
builtin:host.cpu.load15 |
number | ||
builtin:host.memory.total |
byte | ||
builtin:host.memory.used |
byte | ||
builtin:host.memory.reclaimable |
byte | ||
builtin:host.memory.kernel |
byte | Windows kernel memory is the portion of system memory reserved for the Windows operating system kernel and its components, including device drivers and core system services. | |
builtin:host.memory.swap.total |
byte | ||
builtin:host.memory.swap.used |
byte | ||
builtin:host.memory.pgfault |
number | ||
builtin:host.disk.total |
byte | ||
builtin:host.disk.used |
byte | ||
builtin:host.network.traffic.in |
bit | ||
builtin:host.network.traffic.out |
bit | ||
builtin:host.network.packets.in |
number | ||
builtin:host.network.packets.out |
number | ||
builtin:host.network.errors.in |
number | ||
builtin:host.network.errors.out |
number | ||
builtin:host.network.drops.in |
number | ||
builtin:host.network.drops.out |
number |