Metrics Reference

Understanding metrics collected by DC/OS

Mesosphere DC/OS collects basic system metrics–such as CPU and memory–for nodes and containers automatically. Mesosphere DC/OS also collects metadata about the different categories of metrics. For more information about the metadata metrics, see Dimensions.

You should note that automatically-collected metrics are only available for containers that provide endpoint statistics. For example, Docker containers do not provide networking data for DC/OS to consume, so the networking metrics that are available for UCR containers are not available for Docker containers.


CPU and memory metrics

Metric Description
cpu.idle Percentage of CPUs idle.
cpu.system Percentage of system used. Percentage of CPUs used.
cpu.user Percentage of CPU used by the user.
cpu.wait Percentage idle while waiting for an operation to complete.
load.1min Load average for the past minute.
load.5min Load average for the past 5 minutes.
load.15min Load average for the past 15 minutes.
memory.buffers Number of memory buffers.
memory.cached Amount of cached memory. Amount of free memory in bytes. Total memory in bytes.
process.count Number of processes that are running. Amount of free swap space. Total swap space.
swap.used Amount of swap space used.
system.uptime The system uptime.

File system metrics

Metric Description Amount of available capacity in bytes. Total capacity in bytes.
filesystem.capacity.used Capacity used in bytes. Amount of available inodes in bytes. Total inodes in bytes.
filesystem.inode.used Inodes used in bytes.

NOTE: The tag path is automatically populated based on the mount path of the local filesystem (for example, /, /boot, etc).

Network interface metrics

Metric Description Number of bytes downloaded. Number of downloaded bytes dropped. Number of downloaded bytes in error. Number of packets downloaded.
network.out Number of bytes uploaded.
network.out.dropped Number of uploaded bytes dropped.
network.out.errors Number of uploaded bytes in error.
network.out.packets Number of packets uploaded.

NOTE: The tag interface is automatically populated based on the type of the network interface (for example, spartan, d-dcos, minuteman, etc).


The following per-container resource utilization metrics are collected.

CPU usage metrics

Metric Description
cpus.limit The number of CPU shares allocated.
cpus.system_time_secs Total CPU time spent in kernel mode in seconds.
cpus.throttled_time_secs Total time, in seconds, that CPU was throttled.
cpus.user_time_secs Total CPU time spent in user mode.

Disk metrics

Metric Description
disk.limit_bytes Hard capacity limit for disk in bytes.
disk.used_bytes Hard capacity used in bytes.

Memory metrics

Metric Description
mem.limit_bytes Hard memory limit for a container.
mem.total_bytes Total memory of a process in RAM (as opposed to in swap).

Network metrics

Metric Description
net.rx.bytes Bytes received.
net.rx.dropped Packets dropped on receive.
net.rx.errors Errors reported on receive.
net.rx.packets Packets received.
net.tx.bytes Bytes sent.
net.tx.dropped Packets dropped on send.
net.tx.errors Errors reported on send.
net.tx.packets Packets sent.


Dimensions are metadata about the metrics. The following table lists the available dimensions and the entities where they appear.

Dimension Description Entity
mesos_id The Mesos ID of the node. node, container
cluster_id The ID of the Mesos cluster. node, container
container_id The ID of the container. metric, container
executor_name The name of the task executor. metric
framework_name The name of the framework. container
hostname The IP address of the node. container, node
labels Key-value pairs describing the metric. container
task_name The task name. container

For more information, see the dcos-metrics repository documentation.