Important: This documentation is about an older version. It's relevant only to the release noted, many of the features and functions have been updated or replaced. Please view the current version.

Documentation Grafana Agent Flow mode Reference Components prometheus.relabel

prometheus.relabel

Prometheus metrics follow the OpenMetrics format. Each time series is uniquely identified by its metric name, plus optional key-value pairs called labels. Each sample represents a datapoint in the time series and contains a value and an optional timestamp.

Copy

<metric name>{<label_1>=<label_val_1>, <label_2>=<label_val_2> ...} <value> [timestamp]

The prometheus.relabel component rewrites the label set of each metric passed along to the exported receiver by applying one or more relabeling rules. If no rules are defined or applicable to some metrics, then those metrics are forwarded as-is to each receiver passed in the component’s arguments. If no labels remain after the relabeling rules are applied, then the metric is dropped.

The most common use of prometheus.relabel is to filter Prometheus metrics or standardize the label set that is passed to one or more downstream receivers. The rule blocks are applied to the label set of each metric in order of their appearance in the configuration file. The configured rules can be retrieved by calling the function in the rules export field.

Multiple prometheus.relabel components can be specified by giving them different labels.

Usage

river Copy

prometheus.relabel "LABEL" {
  forward_to = RECEIVER_LIST

  rule {
    ...
  }

  ...
}

Arguments

The following arguments are supported:

Blocks

The following blocks are supported inside the definition of prometheus.relabel:

rule block

The rule block contains the definition of any relabeling rules that can be applied to an input metric. If more than one rule block is defined, the transformations are applied in top-down order.

The following arguments can be used to configure a rule. All arguments are optional. Omitted fields take their default values.

Here’s a list of the available actions, along with a brief description of their usage.

• replace - Matches regex to the concatenated labels. If there’s a match, it replaces the content of the target_label using the contents of the replacement field.
• keep - Keeps metrics where regex matches the string extracted using the source_labels and separator.
• drop - Drops metrics where regex matches the string extracted using the source_labels and separator.
• hashmod - Hashes the concatenated labels, calculates its modulo modulus and writes the result to the target_label.
• labelmap - Matches regex against all label names. Any labels that match are renamed according to the contents of the replacement field.
• labeldrop - Matches regex against all label names. Any labels that match are removed from the metric’s label set.
• labelkeep - Matches regex against all label names. Any labels that don’t match are removed from the metric’s label set.
• keepequal - Drop targets for which the concatenated source_labels do not match target_label.
• dropequal - Drop targets for which the concatenated source_labels do match target_label.
• lowercase - Sets target_label to the lowercase form of the concatenated source_labels.
• uppercase - Sets target_label to the uppercase form of the concatenated source_labels.

Finally, note that the regex capture groups can be referred to using either the $CAPTURE_GROUP_NUMBER or ${CAPTURE_GROUP_NUMBER} notation.

Exported fields

The following fields are exported and can be referenced by other components:

Component health

prometheus.relabel is only reported as unhealthy if given an invalid configuration. In those cases, exported fields are kept at their last healthy values.

Debug information

prometheus.relabel does not expose any component-specific debug information.

Debug metrics

• agent_prometheus_relabel_metrics_processed (counter): Total number of metrics processed.
• agent_prometheus_relabel_metrics_written (counter): Total number of metrics written.
• agent_prometheus_relabel_cache_misses (counter): Total number of cache misses.
• agent_prometheus_relabel_cache_hits (counter): Total number of cache hits.
• agent_prometheus_relabel_cache_size (gauge): Total size of relabel cache.
• agent_prometheus_fanout_latency (histogram): Write latency for sending to direct and indirect components.
• agent_prometheus_forwarded_samples_total (counter): Total number of samples sent to downstream components.

Example

Let’s create an instance of a see prometheus.relabel component and see how it acts on the following metrics.

river Copy

prometheus.relabel "keep_backend_only" {
  forward_to = [prometheus.remote_write.onprem.receiver]

  rule {
    action        = "replace"
    source_labels = ["__address__", "instance"]
    separator     = "/"
    target_label  = "host"
  }
  rule {
    action        = "keep"
    source_labels = ["app"]
    regex         = "backend"
  }
  rule {
    action = "labeldrop"
    regex  = "instance"
  }
}

Copy

metric_a{__address__ = "localhost", instance = "development", app = "frontend"} 10
metric_a{__address__ = "localhost", instance = "development", app = "backend"}  2
metric_a{__address__ = "cluster_a", instance = "production",  app = "frontend"} 7
metric_a{__address__ = "cluster_a", instance = "production",  app = "backend"}  9
metric_a{__address__ = "cluster_b", instance = "production",  app = "database"} 4

After applying the first rule, the replace action populates a new label named host by concatenating the contents of the __address__ and instance labels, separated by a slash /.

Copy

metric_a{host = "localhost/development", __address__ = "localhost", instance = "development", app = "frontend"} 10
metric_a{host = "localhost/development", __address__ = "localhost", instance = "development", app = "backend"}  2
metric_a{host = "cluster_a/production",  __address__ = "cluster_a", instance = "production",  app = "frontend"} 7
metric_a{host = "cluster_a/production",  __address__ = "cluster_a", instance = "production",  app = "backend"}  9
metric_a{host = "cluster_b/production",  __address__ = "cluster_a", instance = "production",  app = "database"} 4

On the second relabeling rule, the keep action only keeps the metrics whose app label matches regex, dropping everything else, so the list of metrics is trimmed down to:

Copy

metric_a{host = "localhost/development", __address__ = "localhost", instance = "development", app = "backend"}  2
metric_a{host = "cluster_a/production",  __address__ = "cluster_a", instance = "production",  app = "backend"}  9

The third and final relabeling rule which uses the labeldrop action removes the instance label from the set of labels.

So in this case, the initial set of metrics passed to the exported receiver is:

Copy

metric_a{host = "localhost/development", __address__ = "localhost", app = "backend"}  2
metric_a{host = "cluster_a/production",  __address__ = "cluster_a", app = "backend"}  9

The two resulting metrics are then propagated to each receiver defined in the forward_to argument.