Metrics and Labels in Prometheus Format

The Prometheus metrics library lists the metrics in Prometheus format supported by the Sysdig product suite, as well as kube state and cloud provider metrics.

The metrics listed in this section follows the statsd-compatible Sysdig naming convention. To see a mapping between Prometheus notation and Sysdig notation, see Metrics and Label Mapping.

Overview

Each metric in the dictionary has several pieces of metadata listed to provide greater context for how the metric can be used within Sysdig products. An example layout is displayed below:

Metric Name

Metric definition. For some metrics, the equation for how the value is determined is provided.

Metadata

Definition

Metric Type

Metric type determines whether the metric value is a counter metric or a gauge metric. Sysdig Monitor offers two Metric types:

Counter: The metric whose value keeps on increasing and is reliant on previous values. It helps you record how many times something has happened, for example, a user login.

Gauge: Represents a single numerical value that can arbitrarily fluctuate over time. Each value returns an instantaneous measurement, for example, CPU usage.

Value Type

The type of value the metric can have. The possible values are:

  • Percent (%)

  • Byte

  • Date

  • Double

  • Integer (int)

  • relativeTime

  • String

Segment By

The levels within the infrastructure that the metric can be segmented at:

  • Host

  • Container

  • Process

  • Kubernetes

  • Mesos

  • Swarm

  • CloudProvider

Default Time Aggregation

The default time aggregation format for the metric.

Available Time Aggregation Formats

The time aggregation formats the metric can be aggregated by:

  • Average (Avg)

  • Rate

  • Sum

  • Minimum (Min)

  • Maximum (Max)

Default Group Aggregation

The default group aggregation format for the metric.

Available Group Aggregation Formats

The group aggregation formats the metric can be aggregated by:

  • Average (Avg)

  • Sum

  • Minimum (Min)

  • Maximum (Max)

Topics in This Section
Agent

Containers

Metric Labels

File

Due to the nature of container storage with layered and overlay filesystems, segmenting fs.* metrics by container may return unexpected results and should be used with care. Typically they will report statistics of host filesystems mounted to the container, which can report duplication of metrics. They are not designed to report storage usage by container.

Host

JMX/JVM

Kubernetes

Network

Program

Provider

Cost