System Metrics with InstrumentalD

InstrumentalD automatically reports a variety of system metrics for every machine on which it's installed. These metrics will help you find hotspots in your infrastructure, and plan for future infrastructure changes and maintenance - like when you are limited by your CPUs, disks, and other physical quantities.

InstrumentalD reports the following kinds of metrics out of the box:

You can optionally configure InstrumentalD to only collect a subset of these metrics (or none at all.)

CPU Stats

InstrumentalD reports aggregate CPU stats in the following format:

system.<hostname>.cpu.<metric>
All CPU metrics are reported in percent. The individual metrics are:
usage_guest Time spent running a virtual CPU for guest operating systems under the control of the kernel
usage_guest_nice Time spent running a niced guest (virtual CPU for guest operating systems under the control of the Linux kernel)
usage_idle Time spent in the idle task
usage_iowait Time waiting for I/O to complete
usage_irq Time servicing interrupts
usage_nice Time spent in user mode with low priority (nice)
usage_softirq Time servicing softirqs
usage_steal Stolen time, which is the time spent in other operating systems when running in a virtualized environment
usage_system Time spent in system mode
usage_user Time spent in user mode

Disk Stats

InstrumentalD reports disk statistics in the following format:

system.<hostname>.disk.<mount-point>.<metric>

Note: due to a limitation in our metric names, we report / as -

The following stats will be recorded for each mount point:
free Free disk in bytes
total Total size of disk in bytes
used Used disk in bytes
used_percent Percentage of disk in use

Load Stats

InstrumentalD reports load statistics in the following format:

system.<hostname>.load.<metric>
The individual metrics are:
load1 Load average in the past minute
load5 Load average in the past 5 minutes
load15 Load average in the past 15 minutes

Memory Stats

InstrumentalD reports memory statistics in the following format:

system.<hostname>.memory.<metric>
The individual metrics are:
active Memory that is being used by running processes in bytes
available The actual amount of available memory that can be given instantly to processes that request more memory in bytes
available_percent Percent of memory available
buffered Size of in-memory block I/O buffers bytes
cached Size of the page cache in bytes
free Memory not being used at all (zeroed) that is readily available in bytes
inactive Memory that was allocated to processes that are no longer running in bytes
total Total physical memory available in bytes
used Memory in use in bytes
used_percent Percentage of memory in use

Network Stats

InstrumentalD reports network statistics in the following format:

system.<hostname>.network.<network-adapter>.<metric>
The following metrics will be recorded for each network adapter (e.g. eth0 on the system being monitored:
bytes_recv Bytes received
bytes_sent Bytes sent
drop_in Inbound packets dropped
drop_out Outbound packets dropped
err_in Receive errors
err_out Transmit errors
packets_recv Packets received
packets_sent Packets transmitted

Swap Stats

InstrumentalD reports swap statistics in the following format:

system.<hostname>.swap.<metric>
The individual metrics are:
free Swap space not being used that is readily available in bytes
in Pages swapped in, total
out Pages swapped out, total
total Total swap space
used Swap space in use in bytes
used_percent Percent of swap space in use

Suggested Graphs

The following graphs are based on the metrics collected by InstrumentalD, and provide great starting point for monitoring your servers.

Log in or sign up to add them to your project.