Skip to content
share library_books

Operating System Metrics

The agent collects basic system metrics including memory usage, CPU utilization, disk I/O statistics, and more. Most metrics are directly pulled from the procsfs every 10 seconds, although it is possible to alter the collection interval by modifying the st-agent.yml configuration file and specifying the period along with the time unit:

# Metric collection interval for OS, network & container metrics
interval: 10s

The following is the detailed breakdown of every single metric tracked by Sematext Agent. Note that disk and network metrics are reported per device.

Memory

Metric name Description
total memory Represents the overall amount of physical memory available on the system
available memory Designates the available physical memory for processes to allocate
used memory Is the amount of physical memory allocated by processes
free memory Represents the amount of free physical memory
buffered memory The amount of physical memory used for file system buffers where temporary raw blocks are allocated before they are flushed to disk
cache memory Is the amount of physical memory used as cache memory for files read from the disk
total swap memory Represents the total amount of the swap area where inactive memory pages are pushed
free swap memory Is the the amount of free swap memory
swap pages in The count of memory pages pulled from the swap area
swap pages out The count of memory pages swapped to disk

CPU

Metric name Description
user time The time processes have spent executing in user space
sys time The time processes have spent executing in kernel space
idle time The amount of CPU idle time
nice time Is the amount of time processes have been executing with nice priority
wait time Represents the time spent serving I/O requests
irq time Is the time spent serving hardware interrupts
softirq time Designates the time spent serving software interrupts (e.g. deferred tasks scheduled by the kernel)
steal time Is the amount of CPU time desired by a guest virtual machine

Disk

Metric name Description
read bytes Is the number of bytes read from the storage layer
written bytes Is the number of bytes written to the storage layer
free bytes Represents the number of free bytes for the block device
used bytes Included the number of used bytes for the block device

System load

Metric name Description
load Average number of jobs executing in the kernel's run queue
load 5 Average number of jobs executing in the run queue for the last 5 minutes
load 15 Average number of jobs executing in the run queue for the last 15 minutes

Network

Metric name Description
received bytes Received amount of bytes on the network interface
received packets Received amount of packets on the network interface
received errors Received amount of errors on the network interface
dropped ingress packets The amount of dropped inbound packets on the network interface
transmitted bytes Transmitted amount of bytes on the network interface
transmitted packets Transmitted amount of packets on the network interface
transmitted errors Transmitted amount of errors on the network interface
dropped egress packets The amount of dropped outbound packets on the network interface