Docs » SignalFx Smart Agent » Install and configure the SignalFx Smart Agent

Install and configure the SignalFx Smart Agent πŸ”—


The SignalFx Smart Agent is deprecated and will reach end of support on June 30th, 2023. Note that this only affects the agent; Smart Agent receivers bundled in the Splunk Open Telemetry Collector are not deprecated. For details, see the Deprecation Notice.

The SignalFx Smart Agent gathers host performance, application, and service-level metrics from both containerized and non-container environments. This page provides a complete list of Smart Agent resources. Use your browser’s search function to quickly find what you’re looking for.

See SignalFx Smart Agent commands reference for a list of the most commonly used commands for the SignalFx Smart Agent.

Components πŸ”—

The agent has three main components:

  • Observers that discover applications and services running on the host. For a list of supported observers and their configurations, see observer configuration.

  • Monitors that collect metrics, events, and dimension properties from the host and applications. For a list of supported monitors and their configurations, see monitor configuration.

  • The writer that sends the metrics, events, and dimension updates collected by monitors to Splunk Observability Cloud. The writer collects metrics emitted by configured monitors and sends them to Observability Cloud on a regular basis. You can configure writer settings in the configuration schema.

Use cases πŸ”—

The Smart Agent gathers metrics using monitors, including Python-based plugins such as Mongo, Redis, and Docker. See Supported integrations for a list of data source integrations.

Use the Smart Agent to integrate with cloud services, including Amazon Web Services, Microsoft Azure, Google Cloud Platform, and Kubernetes environments. See Connect to your cloud service provider. Next, log in to Observability Cloud to view the incoming metrics in Dashboards in Splunk Observability Cloud and Charts in Splunk Observability Cloud.

Check out the health of your network and nodes using navigators. Some features of Observability Cloud such as related content do not work with the Smart Agent. To learn more, see Use navigators in Splunk Infrastructure Monitoring.

The Smart Agent also supports receiving and sending trace data. See Download traces.

Check exposed ports πŸ”—

Before installing the Smart Agent, check exposed ports to make sure your environment doesn’t have conflicts. You can change the ports in the Smart Agent configuration.

  • Port 8095 is the default port that the internal status server listens on. Configure port 8095 using the internalStatusPort option.

  • Port 9080 is the default port that the server listens on. Configure port 9080 using the listenAddress option. The listenAddress option is a configurable option for the trace-forwarder and signalfx-forwarder monitors.

Install the Smart Agent πŸ”—

There are several options available to install the Smart Agent. Select the option that matches your situation or preference. Each of the following links includes prerequisites, configuration instructions, installation instructions, and instructions for verifying your installation.

The Smart Agent is incompatible on Linux systems with SELinux enabled. Check the documentation for your distribution to learn how to disable SELinux.


To uninstall the Smart Agent, see Uninstall the Smart Agent.

Configure the Smart Agent πŸ”—

You can configure the Smart Agent by editing the agent.yaml file. By default, the configuration is installed at and looked for at /etc/signalfx/agent.yaml on Linux and \ProgramData\SignalFxAgent\agent.yaml on Windows. You can override default locations using the -config command line flag.

The configuration schema includes the options that you can use in the agent.yaml file to control the behavior of your integrations. Example.yaml provides an autogenerated example of a YAML configuration file, with default values where applicable. Remote configuration describes how to configure the Smart Agent from remote sources, such as other files on the file system, or from key-value stores such as etcd.

Capture logs using the Smart Agent πŸ”—

The default log level is info, which logs Smart Agent events without spamming the logs. Most of the info level logs are created upon startup and capture service discovery changes to record messages about routine operations. The debug log level creates verbose log output and should only be used when troubleshooting.

You can change the log level using the logging: {level: info} YAML configuration option. Other valid log levels include warn and error.

The Smart Agent sends logs either as unstructured text (default) or JSON format. You can configure the Smart Agent to send JSON logs using the logging: {format: json} YAML configuration option.

Linux πŸ”—

The Smart Agent supports logging to stdout/stderr, which is generally redirected by the init scripts provided to either a file at /var/log/signalfx-agent.log or to the systemd journal on newer distros.

Windows πŸ”—

On Windows, the Smart Agent logs to the console when executed directly in a shell. If the Smart Agent is configured as a Windows service, log events are logged to the Windows Event Log. Select Start > Administrative Tools > Event Viewer to read logs. Select Windows Logs > Application to see logged events from the Smart Agent service.

Enable proxy support in the Smart Agent πŸ”—

To use an HTTP or HTTPS proxy, set the environment variable HTTP_PROXY and/or HTTPS_PROXY in the container configuration to proxy either protocol. The SignalFx ingest and API servers both use HTTPS. If the NO_PROXY environment variable exists, the Smart Agent automatically appends the local services to the environment variable to not use the proxy.

If the Smart Agent is running as a local service on the host, refer to the host documentation for information on passing environment variables to the Smart Agent service to enable proxy support when the service is started. For example, if the host services are managed by systemd, create the /etc/systemd/system/signalfx-agent.service.d/myproxy.conf file and add the following to the file:


After updating the file, run systemctl daemon-reload and systemctl restart signalfx-agent.service to restart the service with proxy support.

Sys-V based init.d systems: Debian and RHEL πŸ”—

Create the /etc/default/signalfx-agent.yaml file with the following contents:


Smart Agent diagnostics πŸ”—

The Smart Agent serves diagnostic information on an HTTP server at the address configured by the internalStatusHost and internalStatusPort options. Use the signalfx-agent command status to read the server and extract its contents. Use the content to identify and resolve issues with the Smart Agent. The signalfx-agent command also explains how to get further diagnostic information.

Service discovery using the Smart Agent πŸ”—

The Smart Agent includes a comprehensive service discovery feature. This feature allows the Smart Agent to identify each of the services within your environment and automatically configure the appropriate integration plugins from within its bundle. This is particularly valuable in large ephemeral container environments that experience high-churn and dynamic service creation, as new services are automatically discovered, installed, and configured within minutes. However, this capability is extended to non-containerized environments as well. See Service Discovery for more information and configuration options.

Filtering data using the Smart Agent πŸ”—

Filter out certain data points or properties to prevent them from ever leaving the Smart Agent. Filtering can be useful to reduce clutter in charts without having to resort to filtering in the UI.

If possible, it is preferable to prevent the data points and properties you want to omit from being generated by a monitor in the first place, as this reduces CPU and memory usage of the Smart Agent, but sometimes this is not feasible.

See Filtering for more information and configuration options for Smart Agent 4.7.0+. See Legacy Filtering for more information on the old style of filtering, which is deprecated and removed in Smart Agent 5.0+.

Uninstall the Smart Agent πŸ”—

To remove the Smart Agent from your system, follow the steps for each installation method.

Linux (package manager) πŸ”—

  • For Debian-based distributions, including Ubuntu, run the following command:

    sudo dpkg --remove signalfx-agent
  • For Red Hat, CentOS, and other RPM-based installs, run the following command:

    sudo rpm -e signalfx-agent


Configuration files might persist in /etc/signalfx.

Linux (tar file) πŸ”—

To uninstall the Smart Agent, stop the signalfx-agent process and delete the signalfx-agent directory.

Windows (installer) πŸ”—

Uninstall the Smart Agent from Programs and Features in the Control Panel.


Configuration files might persist in \ProgramData\SignalFxAgent.

Windows (ZIP file) πŸ”—

Run the following PowerShell commands to stop and uninstall the signalfx-agent service:

SignalFxAgent\bin\signalfx-agent.exe -service "stop"
SignalFxAgent\bin\signalfx-agent.exe -service "uninstall"

Then, delete the SignalFxAgent folder.

Deployed using kubectl πŸ”—

To delete all Smart Agent resources, run the following command in the directory containing the YAML configuration files:

cat *.yaml | kubectl delete -f -

For more details, see the Kubectl Reference Documentation.

Deployed using Helm πŸ”—

To uninstall the Helm release of the Smart Agent, follow these steps:

Deployed in AWS ECS πŸ”—

To deregister the signalfx-agent task definitions, see the official AWS documentation at

To delete the signalfx-agent service, see Deleting a service at

Frequently asked questions πŸ”—

See Frequently Asked Questions to troubleshoot issues with the Smart Agent.

Migrating from the Smart Agent to the Splunk Distribution of OpenTelemetry Collector πŸ”—

The Splunk Distribution of OpenTelemetry Collector is the next-generation agent and gateway for Splunk Observability products. See the following topics for information on migrating to the Splunk Distribution of OpenTelemetry Collector:

How mapping makes upgrades easier πŸ”—

The Mapping Service is a transition tool that defines equivalencies between legacy collectd (Smart Agent) metric naming and semantic conventions to the OpenTelemetry names and formats for metrics and metric metadata. Mapping supports multiple observers, deployment types, and kinds of metadata.

The Mapping Service enables you to migrate from SignalFx Smart Agent deployments to OpenTelemetry without significantly disrupting the form or content of your existing dashboards and detectors. The Mapping Service also enables you to slowly transition from the Smart Agent to OpenTelemetry across your organization (though you cannot use both agents simultaneously on the same host).

Mapping happens automatically as a background operation, but you can view mapping definitions in Metric Finder or in the Mapping and OTel Transition Impact Report.

The mapping impact report explains how the transition from Smart Agent to OpenTelemetry affects some of the variables and saved filters in the following components:

  • Dashboards

  • Charts

  • Detectors

The mapping impact report also tells you where to find whatever subset of your content calls functions with Smart Agent names, so that you can update that content either by hand or programmatically to complete your transition to open telemetry.

What is flagged for update in translation πŸ”—

The Mapping and OTel Transition Impact Report is specific to your computing environment. The report flags the following items and tells you where to find and update them in your collection of plots, filters, and functions:

  • Wildcards

  • Direct references to Smart Agent metrics

  • Filters that use Smart Agent dimensions

  • Aggregates that use Smart Agent dimensions

The mapping impact report also shows which OpenTelemetry metrics and dimensions work well as replacements for specific Smart Agent metrics and dimensions, with the important exception of wildcards not supported by OpenTelemetry.

You can view and save the mapping impact report even if you opt out of mapping.

About the mapping service transition impact report πŸ”—

To make migration of your data and metadata from dashboards, charts, and detectors into OpenTelemetry (OTel) as seamless as possible, the application’s Mapping Service automatically translates collectd (Smart Agent) conventions into the syntax used by the OpenTelemetry Collector.

Mapping applies OpenTelemetry naming standards to the following components:

  • Metrics and metric time series (MTS)

  • Dimensions

  • Properties

For example, if you track CPU utilization for your Kubernetes pod, your analytics may use the kubernetes.container_cpu_limit value. In that case, the mapping service updates your existing queries to include both legacy semantics and new semantics (such as k8s.container.cpu_limit) joined by an OR clause. The Mapping Service creates equivalencies between your Smart Agent and OTel metric names.

Mapping logic treats any of the names for a metric or property as referring to that same metric or property in OpenTelemetry, without tracking versions.

You can find a table outlining OpenTelemetry values and their legacy equivalents in GitHub at: Legacy to OTel semantics mapping table.

Whether you’re using Smart Agent collection or OTel collection, your original dashboards and detectors function the same way. Infrastructure Navigator views use the mapping service to show both old collection data and new collection data.

If you decide as a Splunk admin to turn off the mapping service, you can still generate and download a Mapping and OTel Transition Impact Report specific to migration for your cloud computing environment.

To access the migration transition impact report, follow these steps:

  1. Log in to Splunk Observability Cloud.

  2. In the left navigation menu, select Settings > Billing and Usage.

  3. Click the View detailed usage reports link.

  4. Select the OpenTelemetry Migration tab.

  5. Click Download to open the report as a comma-separated values file.

Interpreting the mapping impact report πŸ”—

The Mapping and OTel Transition Impact Report summarizes the scope of component name change associated with your migration to open telemetry. It assesses your data set to list the tokens currently used as metric, dimension, property or tag names, and highlights migration rules that could generate conflict between old and new equivalence groups.

The report explains when migration from an old MTS to a new MTS will trigger detectors, and which detectors those are. For example, heartbeat detectors working with un-aggregated MTS are affected by design, but if a heartbeat detector is working with a dimension that continues across the transition to OTel, then the mapping service ensures continuity so that you do not have to restart that detector.

The mapping transition impact report assesses migration effects across three categories:

  • Data object types

  • Team responsibilities

  • Migration mitigation steps you should take

Avoiding unexpected results πŸ”—

Because Mapping Service only renames existing MTS when filtering or grouping requires renaming to conform to OpenTelemetry Collector conventions, correlation across different datasets yields unexpected results when a mapped MTS is correlated with an unmapped MTS. This can happen, for example, when an MTS attempts to correlate with a time-shifted or transformed version of itself.

If you have charts and detectors built from formulas whose terms use different agents, you won’t get the alerts you expect.

To resolve this issue, explicitly filter or group by dimensions so that Mapping Service renames the fields in all the MTS in the job to match the name you specified in the filter or grouping.

Data object type information πŸ”—

The mapping impact report explains migration impacts within your organization to the following object types:

  • Dashboards

  • Charts

  • Detectors

The report shows how many objects of each type are affected, and includes tables that show where to find the affected objects. You can read the report to see, for example, a list of all affected charts on a given dashboard or within a dashboard group.

Team information πŸ”—

The mapping impact report extracts information from your data set about stakeholders, meaning the people who created object types or are affected by changes to them because they’re on email lists of employees to be notified in the event of, for example, a detector being triggered by a critical alert condition.

When applicable, the report shows the names of teams linked to particular detectors. The report also identifies people or teams linked to particular dashboard groups.

Migration mitigation steps πŸ”—

The mapping impact report explains what effect migration will have on the content highlighted in it, so that you can modify that content as needed to ensure a smoother transition.

Flagged items that need to be modified include the following (as listed in the report):

  • Wildcards used in a plot, filter, or function.

  • Direct references to Smart Agent metrics.

  • Filters that use Smart Agent dimensions.

  • Aggregates that use Smart Agent dimensions.

While the mapping impact report highlights items that need revising because they use legacy syntax or conventions, it also pairs those items with the OTel-based metrics and dimensions that you can use as substitutes for them.

In other words, the mapping transition impact report helps take guesswork out of migration tasks.