statsd_exporter_config
Caution
Grafana Alloy is the new name for our distribution of the OTel collector. Grafana Agent has been deprecated and is in Long-Term Support (LTS) through October 31, 2025. Grafana Agent will reach an End-of-Life (EOL) on November 1, 2025. Read more about why we recommend migrating to Grafana Alloy.
Documentation
Grafana Agent
Static mode
Configure static mode
integrations_config
statsd_exporter_config
Open source
statsd_exporter_config
The statsd_exporter_config
block configures the statsd_exporter
integration, which is an embedded version of
statsd_exporter
. This allows
for the collection of statsd metrics and exposing them as Prometheus metrics.
Full reference of options:
yaml
# Enables the statsd_exporter integration, allowing the Agent to automatically
# collect system metrics from the configured statsd server address
[enabled: <boolean> | default = false]
# Sets an explicit value for the instance label when the integration is
# self-scraped. Overrides inferred values.
#
# The default value for this integration is inferred from the agent hostname
# and HTTP listen port, delimited by a colon.
[instance: <string>]
# Automatically collect metrics from this integration. If disabled,
# the statsd_exporter integration will be run but not scraped and thus not
# remote-written. Metrics for the integration will be exposed at
# /integrations/statsd_exporter/metrics and can be scraped by an external
# process.
[scrape_integration: <boolean> | default = <integrations_config.scrape_integrations>]
# How often should the metrics be collected? Defaults to
# prometheus.global.scrape_interval.
[scrape_interval: <duration> | default = <global_config.scrape_interval>]
# The timeout before considering the scrape a failure. Defaults to
# prometheus.global.scrape_timeout.
[scrape_timeout: <duration> | default = <global_config.scrape_timeout>]
# Allows for relabeling labels on the target.
relabel_configs:
[- <relabel_config> ... ]
# Relabel metrics coming from the integration, allowing to drop series
# from the integration that you don't care about.
metric_relabel_configs:
[ - <relabel_config> ... ]
# How frequent to truncate the WAL for this integration.
[wal_truncate_frequency: <duration> | default = "60m"]
#
# Exporter-specific configuration options
#
# The UDP address on which to receive statsd metric lines. An empty string
# will disable UDP collection.
[listen_udp: <string> | default = ":9125"]
# The TCP address on which to receive statsd metric lines. An empty string
# will disable TCP collection.
[listen_tcp: <string> | default = ":9125"]
# The Unixgram socket path to receive statsd metric lines. An empty string
# will disable unixgram collection.
[listen_unixgram: <string> | default = ""]
# The permission mode of the unixgram socket, when enabled.
[unix_socket_mode: <string> | default = "755"]
# An optional mapping config that can translate dot-separated StatsD metrics
# into labeled Prometheus metrics. For full instructions on how to write this
# object, see the official documentation from the statsd_exporter:
#
# https://github.com/prometheus/statsd_exporter#metric-mapping-and-configuration
#
# Note that a SIGHUP will not reload this config.
[mapping_config: <statsd_exporter.mapping_config>]
# Size (in bytes) of the operating system's transmit read buffer associated
# with the UDP or unixgram connection. Please make sure the kernel parameters
# net.core.rmem_max is set to a value greater than the value specified.
[read_buffer: <int> | default = 0]
# Maximum size of your metric mapping cache. Relies on least recently used
# replacement policy if max size is reached.
[cache_size: <int> | default = 1000]
# Metric mapping cache type. Valid values are "lru" and "random".
[cache_type: <string> | default = "lru"]
# Size of internal queue for processing events.
[event_queue_size: <int> | default = 10000]
# Number of events to hold in queue before flushing.
[event_flush_threshold: <int> | default = 1000]
# Number of events to hold in queue before flushing.
[event_flush_interval: <duration> | default = "200ms"]
# Parse DogStatsd style tags.
[parse_dogstatsd_tags: <bool> | default = true]
# Parse InfluxDB style tags.
[parse_influxdb_tags: <bool> | default = true]
# Parse Librato style tags.
[parse_librato_tags: <bool> | default = true]
# Parse SignalFX style tags.
[parse_signalfx_tags: <bool> | default = true]
# Optional: Relay address configuration. This setting, if provided,
# specifies the destination to forward your metrics.
# Note that it must be a UDP endpoint in the format 'host:port'.
[relay_addr: <string>]
# Maximum relay output packet length to avoid fragmentation.
[relay_packet_length: <int> | default = 1400]
Was this page helpful?
Related documentation
Related resources from Grafana Labs
Additional helpful documentation, links, and articles:
5 Mar

Getting started with managing your metrics, logs, and traces using Grafana
In this webinar, we’ll demo how to get started using the LGTM Stack: Loki for logs, Grafana for visualization, Tempo for traces, and Mimir for metrics.
12 Mar

Intro to Application and Frontend Observability with Grafana Cloud
In this webinar, we will introduce you to two of our latest opinionated solutions in Grafana Cloud that resolve issues faster with unified observability.
2 Apr

How to get started with OpenTelemetry and Grafana
This webinar covers the essentials of instrumentation and demonstrates how to visualize OpenTelemetry data seamlessly within Grafana and Grafana Cloud.