Span Metrics

The Span Metrics connector consumes spans from a traces pipeline and generates Request, Error, and Duration (R.E.D) metrics from them, feeding the resulting metrics into a metrics pipeline. It is built on the OpenTelemetry spanmetricsconnector and lets you derive request rate, error rate, and latency metrics from existing trace data without instrumenting code for metrics separately.

Supported Telemetry Types

Input	Output
Traces	Metrics

The connector receives trace spans from one pipeline, converts them into call-count and duration metrics, and emits those metrics into a metrics pipeline.

Generated Metrics

The connector produces the following metrics (prefixed by the namespace, default traces.span.metrics):

Metric	Type	Description
calls	Sum	Count of spans. Error counts are derived via the status.code attribute.
duration	Histogram	Distribution of span durations.
events	Sum	Count of span events (optional, disabled by default).

Every metric includes these default attributes, unless removed with Exclude Default Attributes:

Attribute	Description
service.name	The service that produced the span.
span.name	The operation name.
span.kind	The span kind (client, server, internal, producer, consumer).
status.code	The span status code (Ok, Error, Unset).

Configuration

Basic Configuration

General Settings

Parameter	Type	Required	Default	Description
Metric Namespace	String	No	(empty)	Prefix for generated metric names. Leave empty for default (traces.span.metrics).
Aggregation Temporality	Enum: AGGREGATION_TEMPORALITY_CUMULATIVE, AGGREGATION_TEMPORALITY_DELTA	No	Cumulative	How metrics are aggregated over time.
Metrics Flush Interval	Duration	No	5s	How often to flush generated metrics. Examples: 15s, 1m, 60s.

Attributes

Parameter	Type	Required	Default	Description
Additional Attributes	Map	No	(empty)	Span or resource attributes to include as metric dimensions on all generated metrics. Map attribute name to a default value (leave value empty if the attribute should be omitted when missing).
Exclude Default Attributes	Strings	No	(empty)	Default attributes to exclude from metrics. Default attributes are service.name, span.name, span.kind, status.code.
Calls Metric Attributes	Map	No	(empty)	Additional attributes for the calls metric only, on top of the common attributes above.

Duration Histogram

Parameter	Type	Required	Default	Description
Enable Duration Histogram	Boolean	No	true	Generate a duration histogram metric. When disabled, only call count metrics are produced.
Histogram Type	Enum: explicit, exponential	No	explicit	Explicit uses fixed bucket boundaries. Exponential uses dynamic boundaries based on scale. Shown when Enable Duration Histogram is true.
Duration Unit	Enum: ms, s	No	ms	Unit for duration measurements in the histogram. Shown when Enable Duration Histogram is true.
Bucket Boundaries	Strings	No	(empty)	Histogram bucket boundaries with units. Leave empty for defaults.1 Shown when Enable Duration Histogram is true and Histogram Type is explicit.
Exponential Max Size	Integer	No	160	Maximum number of buckets per positive or negative range for exponential histogram. Shown when Enable Duration Histogram is true and Histogram Type is exponential.
Duration Histogram Attributes	Map	No	(empty)	Additional attributes for the duration histogram metric only, on top of the common attributes above. Shown when Enable Duration Histogram is true.

1. Defaults: 2ms, 4ms, 6ms, 8ms, 10ms, 50ms, 100ms, 200ms, 400ms, 800ms, 1s, 1400ms, 2s, 5s, 10s, 15s.

Advanced

These parameters are available under the advanced configuration toggle.

Parameter	Type	Required	Default	Description
Enable Exemplars	Boolean	No	false	Attach trace exemplars to histogram metrics, allowing correlation from metrics back to traces.
Max Exemplars Per Data Point	Integer	No	5	Maximum number of exemplars to attach per histogram data point. Shown when Enable Exemplars is true.
Enable Events Metric	Boolean	No	false	Generate a metric counting span events.
Event Attributes	Map	Yes	(empty)	Span event attributes to include as dimensions on the events metric. At least one is required. Map attribute name to a default value (leave value empty if the attribute should be omitted when missing). Shown when Enable Events Metric is true.
Metrics Expiration	Duration	No	0	Time after which metrics stop being exported if no new spans arrive. 0 means never expire.
Resource Metrics Cache Size	Integer	No	1000	Size of the cache holding metrics for a service.
Metric Timestamp Cache Size	Integer	No	1000	Cache size for tracking metric timestamps (used with delta temporality).
Aggregation Cardinality Limit	Integer	No	0	Maximum unique dimension combinations. 0 means unlimited.
Include Instrumentation Scope	Strings	No	(empty)	Only process spans from these instrumentation scope names. If empty, all scopes are included.
Resource Metrics Key Attributes	Strings	No	(empty)	Resource attributes used to build the resource metrics map key. If empty, all resource attributes are used.
Add Resource Attributes	Boolean	No	true	Include resource attributes on generated metrics. Has no effect unless the upstream feature gate connector.spanmetrics.excludeResourceMetrics is enabled.

Examples

HTTP latency with custom buckets and exemplars

This example generates R.E.D metrics under a custom namespace, adds the HTTP request method as a dimension, sets explicit histogram buckets tuned for HTTP latency, and enables exemplars so metric data points link back to the originating traces.

• Metric Namespace: http
• Metrics Flush Interval: 15s
• Additional Attributes: http.request.method (no default value)
• Enable Duration Histogram: true
• Histogram Type: explicit
• Bucket Boundaries: 5ms, 10ms, 25ms, 50ms, 100ms, 250ms, 500ms, 1s, 2s, 5s
• Enable Exemplars: true (advanced)
• Max Exemplars Per Data Point: 5 (advanced)

With exemplars enabled, clicking a metric data point in a compatible backend (for example Grafana with Tempo) links directly to a trace that produced it.

Configuration Tips

• Metrics are emitted on a timer. After adding the connector, wait at least one Metrics Flush Interval before expecting metrics to appear.
• The span.name attribute is included by default. Non-normalized span names (such as GET /user/12345 instead of GET /user/{id}) can explode metric cardinality. Normalize span names with a Transform processor upstream, or set an Aggregation Cardinality Limit as a safety net.
• When events are enabled, Event Attributes is required and must contain at least one attribute, otherwise the configuration is invalid.

Troubleshooting

No metrics are produced

Symptoms: a metrics destination downstream of the connector receives nothing.

Solutions:

1. Confirm the connector sits between a Traces source pipeline and a Metrics destination pipeline, and that spans are actually flowing in.
2. Wait at least one Metrics Flush Interval (default 5s) before expecting the first metrics.

Metric cardinality is too high

Symptoms: backend storage or query cost spikes after enabling the connector.

Solutions:

1. Normalize span names upstream and drop high-cardinality Additional Attributes.
2. Set an Aggregation Cardinality Limit to cap the number of unique dimension combinations.

Standalone Connector

apiVersion: bindplane.observiq.com/v1
kind: Connector
metadata:
  name: span-metrics
spec:
  type: spanmetrics
  parameters:
    - name: namespace
      value: http
    - name: aggregation_temporality
      value: AGGREGATION_TEMPORALITY_CUMULATIVE
    - name: metrics_flush_interval
      value: 15s
    - name: dimensions
      value:
        http.request.method: ""
    - name: histogram_enable
      value: true
    - name: histogram_type
      value: explicit
    - name: histogram_unit
      value: ms
    - name: explicit_buckets
      value:
        - 5ms
        - 10ms
        - 25ms
        - 50ms
        - 100ms
        - 250ms
        - 500ms
        - 1s
        - 2s
        - 5s
    - name: exemplars_enabled
      value: true
    - name: exemplars_max_per_datapoint
      value: 5

Related Resources

• OpenTelemetry spanmetrics connector reference

Bindplane Resources

• Count
• Signal to Metrics
• Routing