Docs » Get started with the Splunk Distribution of the OpenTelemetry Collector » Collector components » SignalFx exporter

SignalFx exporter ๐Ÿ”—


The SignalFx exporter creates and excludes metrics by default. Read on to understand which metrics are created, which ones are filtered out, and learn how to modify this behavior.

The SignalFx exporter is a native OTel component that allows the OpenTelemetry Collector to send metrics and events to SignalFx endpoints. The supported pipeline types are traces, metrics, and logs. See Process your data with pipelines for more information.

While the SignalFx Smart Agent has reached End of Support, OTel native components such as the Smart Agent receiver, the SignalFx receiver, and the SignalFx exporter are available and supported. For information on the receivers, see Smart Agent receiver: and SignalFx receiver.

Get started ๐Ÿ”—


This component is included in the default configuration of the Splunk Distribution of the OpenTelemetry Collector when deploying in host monitoring (agent) mode. See Collector deployment modes for more information.

For details about the default configuration, see Configure the Collector for Kubernetes with Helm, Collector for Linux default configuration, or Collector for Windows default configuration. You can customize your configuration any time as explained in this document.

By default, the Splunk Distribution of OpenTelemetry Collector includes the SignalFx exporter in the traces, metrics, and logs/signalfx pipelines.

Sample configurations ๐Ÿ”—

The following example shows the default configuration of SignalFx exporter for metrics and events ingest, as well as trace and metrics correlation:

# Metrics + Events
  access_token: "${SPLUNK_ACCESS_TOKEN}"
  api_url: "${SPLUNK_API_URL}"
  ingest_url: "${SPLUNK_INGEST_URL}"
  # Use instead when sending to gateway (http forwarder extension ingress endpoint)
  #api_url: http://${SPLUNK_GATEWAY_URL}:6060
  #ingest_url: http://${SPLUNK_GATEWAY_URL}:9943
  sync_host_metadata: true

When adding the SignalFx exporter, configure both the metrics and logs pipelines. Make sure to also add the SignalFx receiver as in the following example:

      receivers: [signalfx]
      processors: [memory_limiter, batch, resourcedetection]
      exporters: [signalfx]
      receivers: [signalfx]
      processors: [memory_limiter, batch, resourcedetection]
      exporters: [signalfx]

Default metric filters ๐Ÿ”—

To prevent unwanted custom metrics, the SignalFx exporter excludes a number of metrics by default. See List of metrics excluded by default for more information.

To override default exclusions and include metrics manually, use the include_metrics option. For example:

      - metric_names: [cpu.interrupt, cpu.user, cpu.system]
      - metric_name: system.cpu.time
          state: [interrupt, user, system]

The following example instructs the exporter to send only the cpu.interrupt metric with a cpu dimension value and both per core and aggregate cpu.idle metrics:

      - metric_name: "cpu.idle"
      - metric_name: "cpu.interrupt"
          cpu: ["*"]

List of metrics excluded by default ๐Ÿ”—

Metrics excluded by default by the SignalFx exporter are listed in the default_metrics.go file. The following snippet shows the latest version of the list:

# DefaultExcludeMetricsYaml holds a list of hard coded metrics that's added to the
# exclude list from the config. It includes non-default metrics collected by
# receivers. This list is determined by categorization of metrics in the SignalFx
# Agent. Metrics in the OpenTelemetry convention that have equivalents in the
# SignalFx Agent that are categorized as non-default are also included in this list.


# Metrics in SignalFx Agent Format
- metric_names:
  # CPU metrics.
  - cpu.interrupt
  - cpu.nice
  - cpu.softirq
  - cpu.steal
  - cpu.system
  - cpu.user
  - cpu.utilization_per_core
  - cpu.wait

  # Disk-IO metrics
  - disk_ops.pending

  # Virtual memory metrics
  - vmpage_io.memory.out

# Metrics in OpenTelemetry Convention

# CPU Metrics
- metric_name: system.cpu.time
    state: [idle, interrupt, nice, softirq, steal, system, user, wait]

- metric_name: cpu.idle
    cpu: ["*"]

# Memory metrics
- metric_name: system.memory.usage
    state: [inactive]

# Filesystem metrics
- metric_name: system.filesystem.usage
    state: [reserved]
- metric_name: system.filesystem.inodes.usage

# Disk-IO metrics
- metric_names:
  - system.disk.merged
  - system.disk.time
  - system.disk.io_time
  - system.disk.operation_time
  - system.disk.pending_operations
  - system.disk.weighted_io_time

# Network-IO metrics
- metric_names:

# Processes metrics
- metric_names:
  - system.processes.count
  - system.processes.created

# Virtual memory metrics
- metric_names:
  - system.paging.faults
  - system.paging.usage
- metric_name: system.paging.operations
    type: [minor]

 k8s metrics
- metric_names:
  - k8s.cronjob.active_jobs
  - k8s.job.active_pods
  - k8s.job.desired_successful_pods
  - k8s.job.failed_pods
  - k8s.job.max_parallel_pods
  - k8s.job.successful_pods
  - k8s.statefulset.desired_pods
  - k8s.statefulset.current_pods
  - k8s.statefulset.ready_pods
  - k8s.statefulset.updated_pods
  - k8s.hpa.max_replicas
  - k8s.hpa.min_replicas
  - k8s.hpa.current_replicas
  - k8s.hpa.desired_replicas

  # matches all container limit metrics but k8s.container.cpu_limit and k8s.container.memory_limit
  - /^k8s\.container\..+_limit$/
  - '!k8s.container.memory_limit'
 - '!k8s.container.cpu_limit'

  # matches all container request metrics but k8s.container.cpu_request and k8s.container.memory_request
  - /^k8s\.container\..+_request$/
  - '!k8s.container.memory_request'
  - '!k8s.container.cpu_request'

  # matches any node condition but k8s.node.condition_ready
  - /^k8s\.node\.condition_.+$/
  - '!k8s.node.condition_ready'

  # kubelet metrics
  # matches (container|k8s.node|k8s.pod).memory...
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.memory\.available$/
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.memory\.major_page_faults$/
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.memory\.page_faults$/
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.memory\.rss$/
  - /^(?i:(k8s\.node)|(k8s\.pod))\.memory\.usage$/
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.memory\.working_set$/

  # matches (k8s.node|k8s.pod).filesystem...
  - /^k8s\.(?i:(node)|(pod))\.filesystem\.available$/
  - /^k8s\.(?i:(node)|(pod))\.filesystem\.capacity$/
  - /^k8s\.(?i:(node)|(pod))\.filesystem\.usage$/

  # matches (k8s.node|k8s.pod).cpu.time
  - /^k8s\.(?i:(node)|(pod))\.cpu\.time$/

  # matches (container|k8s.node|k8s.pod).cpu.utilization
  - /^(?i:(container)|(k8s\.node)|(k8s\.pod))\.cpu\.utilization$/

  # matches and
  - /^k8s\.node\.network\.(?:(io)|(errors))$/

  # matches k8s.volume.inodes, k8s.volume.inodes and k8s.volume.inodes.used
  - /^k8s\.volume\.inodes(\.free|\.used)*$/

Filter metrics using service or environment ๐Ÿ”—

The SignalFx exporter correlates the traces it receives to metrics. When the exporter detects a new service or environment, it associates the source (for example, a host or a pod) to that service or environment in Splunk Observability Cloud, and identifies them using sf_service and sf_environment. You can then filter those metrics based on the trace service and environment.


You need to send traces using Splunk APM exporter to see them in Splunk Observability Cloud.

Use the correlation setting to control the syncing of service and environment properties onto dimensions. It has the following options:

  • endpoint: Required. The base URL for API requests, such as Defaults to api_url or https://api.{realm}

  • timeout: Timeout for every attempt to send data to the backend. 5 seconds by default.

  • stale_service_timeout: How long to wait after a spanโ€™s service name is last seen before uncorrelating it. 5 minutes by default.

  • max_requests: Maximum HTTP requests to be made in parallel. 20 by default.

  • max_buffered: Maximum number of correlation updates that can be buffered before updates are dropped. 10,000 by default.

  • max_retries: Maximum number of retries that will be made for failed correlation updates. 2 by default.

  • log_updates: Whether or not to log correlation updates to dimensions, at DEBUG level. false by default.

  • retry_delay: How long to wait between retries. 30 seconds by default.

  • cleanup_interval: How frequently to purge duplicate requests. 1 minute by default.

  • sync_attributes : Map containing key of the attribute to read from spans to sync to dimensions specified as the value. Defaults to {"k8s.pod.uid": "k8s.pod.uid", "": ""}.

See more options in the Settings section.

Translation rules and metric transformations ๐Ÿ”—

Use the translation_rules field to transform metrics or produce custom metrics by copying, calculating, or aggregating other metric values without requiring an additional processor.

Translation rules currently allow the following actions:

  • aggregate_metric: Aggregates a metric through removal of specified dimensions.

  • calculate_new_metric: Creates a new metric via operating on two consistuent ones.

  • convert_values: Convert float values to int or int to float for specified metric names.

  • copy_metrics: Creates a new metric as a copy of another.

  • delta_metric: Creates a new delta metric for a specified non-delta one.

  • divide_int: Scales a metricโ€™s integer value by a given factor.

  • drop_dimensions: Drops dimensions for specified metrics, or globally.

  • drop_metrics: Drops all metrics with a given name.

  • multiply_float: Scales a metricโ€™s float value by a given float factor.

  • multiply_int: Scales a metricโ€™s int value by a given int factor.

  • rename_dimension_keys: Renames dimensions for specified metrics, or globally.

  • rename_metrics: Replaces a given metric name with specified one.

  • split_metric: Splits a given metric into multiple new ones for a specified dimension.

Default translation rules and generated metrics ๐Ÿ”—

The SignalFx exporter uses the translation rules defined in translation/constants.go by default.

The default rules create metrics which are reported directly to Infrastructure Monitoring. If you want to change any of their attributes or values, you need to either modify the translation rules or their constituent host metrics.

By default, the SignalFx exporter creates the following aggregated metrics from the Host metrics receiver:

  • cpu.idle

  • cpu.interrupt

  • cpu.nice

  • cpu.num_processors

  • cpu.softirq

  • cpu.steal

  • cpu.system

  • cpu.user

  • cpu.utilization

  • cpu.utilization_per_core

  • cpu.wait

  • disk.summary_utilization

  • disk.utilization

  • disk_ops.pending



  • memory.utilization


  • process.cpu_time_seconds






  • vmpage_io.memory.out


  • vmpage_io.swap.out

In addition to the aggregated metrics, the default rules make available the following โ€œper coreโ€ custom hostmetrics. The CPU number is assigned to the dimension cpu:

  • cpu.interrupt

  • cpu.nice

  • cpu.softirq

  • cpu.steal

  • cpu.system

  • cpu.user

  • cpu.wait

Drop histogram metrics ๐Ÿ”—

In case of high cardinality metrics, dropping histogram buckets might be useful. To drop histogram metrics, set drop_histogram_buckets to true.

When drop_histogram_buckets is activated, histogram buckets are dropped instead of being translated to datapoints with the _bucket suffix. Only datapoints with _sum, _count, _min, and _max are sent through the exporter.

Settings ๐Ÿ”—

The following table shows the configuration options for the SignalFx exporter:


Use the access_token_passthrough setting if youโ€™re using a SignalFx receiver with the same setting. Only use the SignalFx receiver with the SignalFx exporter when activating this setting.

Troubleshooting ๐Ÿ”—

If you are a Splunk Observability Cloud customer and are not able to see your data in Splunk Observability Cloud, you can get help in the following ways.

Available to Splunk Observability Cloud customers

Available to prospective customers and free trial users

  • Ask a question and get answers through community support at Splunk Answers .

  • Join the Splunk #observability user group Slack channel to communicate with customers, partners, and Splunk employees worldwide. To join, see Chat groups in the Get Started with Splunk Community manual.

To learn about even more support options, see Splunk Customer Success .