collectd-signalfx-metadata.md

collectd/signalfx-metadata

Collectd Python plugin that aggregates various metrics from other collectd plugins. It also sends host metadata to SignalFx through specially formatted events, and sends active process ("top") lists on a periodic basis.

See Python plugin code and Integrations docs.

Monitor Type: collectd/signalfx-metadata

Monitor Source Code

Accepts Endpoints: No

Multiple Instances Allowed: No

Configuration

Config option	Required	Type	Description
writeServerURL	no	string
procFSPath	no	string	(Deprecated) Please set the agent configuration procPath instead of this monitor configuration option. The path to the proc filesystem. Useful to override in containerized environments.
etcPath	no	string	(Deprecated) Please set the agent configuration etcPath instead of this monitor configuration option. The path to the main host config dir. Useful to override in containerized environments.
perCoreCPUUtil	no	bool	Collect the cpu utilization per core, reported as cpu.utilization_per_core. (default: false)
persistencePath	no	string	A directory where the metadata plugin can persist the history of successful host metadata syncs so that host metadata is not sent redundantly. (default: /var/run/signalfx-agent)
omitProcessInfo	no	bool	If true, process "top" information will not be sent. This can be useful if you have an extremely high number of processes and performance of the plugin is poor. (default: false)
dogStatsDPort	no	unsigned integer	Set this to a non-zero value to enable the DogStatsD listener as part of this monitor. The listener will accept metrics on the DogStatsD format, and sends them as SignalFx datapoints to our backend. Setting to a value setting the DogStatsDPort to 0 will result in a random port assignment. Note: The listener emits directly to SignalFx and will not be subject to filters configured with the SignalFx Smart Agent. Internal stats about the SignalFx Smart Agent will not reflect datapoints set through the DogStatsD listener
token	no	string	This is only required when running the DogStatsD listener. Set this to your SignalFx access token.
dogStatsDIP	no	string	Optionally override the default ip that the DogStatsD listener listens on. (default: "0.0.0.0")
ingestEndpoint	no	string	This is optional only used when running the DogStatsD listener. By default the DogStatsD listener will emit to SignalFx Ingest. (default: "https://ingest.signalfx.com")
verbose	no	bool	Set this to enable verbose logging from the monitor (default: false)

Metrics

The following table lists the metrics available for this monitor. Metrics that are marked as Included are standard metrics and are monitored by default.

Name	Type	Included	Description
cpu.utilization	gauge	✔	Percent of CPU used on this host.
cpu.utilization_per_core	gauge		Percent of CPU used on each core. perCoreCPUUtil config must be set to true.
disk.summary_utilization	gauge	✔	Percent of disk space utilized on all volumes on this host.
disk.utilization	gauge	✔	Percent of disk used on this volume.
disk_ops.total	cumulative	✔	Total number of disk read and write operations on this host.
memory.utilization	gauge	✔	Percent of memory in use on this host.
network.total	cumulative	✔	Total amount of inbound and outbound network traffic on this host, in bytes.

To specify custom metrics you want to monitor, add a metricsToInclude filter to the agent configuration, as shown in the code snippet below. The snippet lists all available custom metrics. You can copy and paste the snippet into your configuration file, then delete any custom metrics that you do not want sent.

Note that some of the custom metrics require you to set a flag as well as add them to the list. Check the monitor configuration file to see if a flag is required for gathering additional metrics.

metricsToInclude:
  - metricNames:
    - cpu.utilization_per_core
    monitorType: collectd/signalfx-metadata