Signal to Metrics

Signal to Metrics produces new metrics from incoming logs, metric datapoints, or trace spans. Each metric is defined with an OTTL value expression, an optional filter condition, and a set of attributes to carry through. It bridges pipelines by consuming one or more signal types from a source pipeline and emitting the generated metrics into a metrics pipeline.

Supported Telemetry Types

Input	Output
Logs, Metrics, Traces	Metrics

The connector reads the input signal types you enable and converts matching records into metric datapoints. Whatever the input signal, the output is always Metrics.

Configuration

Basic Configuration

Parameter	Type	Required	Default	Description
Telemetry Types	Telemetry Selector	Yes	Logs, Metrics, Traces	Which input signal types to generate metrics from. At least one must be selected.
Metric Definitions	Signal to Metrics	Yes	(empty)	One or more metrics to generate. At least one metric must be defined. See the fields below.

Metric Definition fields

Each entry under Metric Definitions describes one output metric.

Field	Description
Signal Type	The input signal this metric reads from: logs, datapoints (metrics), or spans (traces). Must be one of the enabled Telemetry Types.
Name	The output metric name, for example http.request.count.
Description	Optional description of the metric.
Unit	Optional unit, for example ms, By, or 1.
Type	Sum, Gauge, Histogram, or Exponential Histogram.
Value	OTTL expression returning the metric value, for example 1 to count, or Double(attributes["duration_ms"]).
Condition	Optional OTTL condition; only matching records generate this metric.
Buckets	Explicit bucket boundaries. Histogram type only.
Max Size	Maximum bucket count. Exponential Histogram type only.
Count	Optional OTTL expression for the datapoint count. Histogram and Exponential Histogram types only.
Attributes	Signal attributes to include on the output metric. See below.
Resource Attributes	Resource attributes to include on the output metric. See below.

Metric types

Type	Behavior
Sum	Aggregates numeric values. Use for counters and cumulative totals.
Gauge	Records the last observed value. Use for current-state measurements.
Histogram	Distributes values across explicit Buckets.
Exponential Histogram	Distributes values across dynamically sized buckets, bounded by Max Size.

Attributes

Choose which attributes from the source signal to copy onto the output metric. If no attributes are defined, the output metric carries no attributes.

Field	Description
Key	The attribute key to include.
Behavior	Required (default): records missing this attribute are skipped. Optional: included when present, omitted when absent. Has Default: a default value is used when the attribute is missing.
Default Value	The value to substitute when the attribute is missing. Only with Has Default.

Resource Attributes

Choose which resource attributes to copy onto the output metric. If no resource attributes are defined, all incoming resource attributes are included.

Field	Description
Key	The resource attribute key to include.
Behavior	Has Default substitutes the default value when missing. (Required is not available for resource attributes.)
Default Value	The value to substitute when the resource attribute is missing. Only with Has Default.

Examples

Latency histogram from HTTP server spans

This metric definition reads trace spans and emits a millisecond-bucketed latency histogram, keeping only server spans and grouping by route and method.

• Signal Type: spans
• Name: http.server.duration
• Unit: ms
• Type: Histogram
• Value: Double(attributes["http.server.request.duration"])
• Condition: kind == SPAN_KIND_SERVER
• Buckets: [5, 10, 25, 50, 100, 250, 500, 1000, 2500, 5000, 10000]
• Attributes: http.route (Optional), http.request.method (Required)

Count log records by severity

A simpler logs-based definition counts every log record and groups by severity.

• Signal Type: logs
• Name: log.record.count
• Type: Sum
• Value: 1
• Attributes: severity (Optional)

Count spans accurately under sampling

For sampled traces, AdjustedCount() scales each span back up to its pre-sampling weight so counts stay accurate.

• Signal Type: spans
• Name: span.count
• Type: Sum
• Value: Int(AdjustedCount())

Configuration Tips

• Each metric definition's Signal Type must be one of the enabled Telemetry Types, otherwise that definition produces nothing.
• Use Value: 1 with a Sum type to turn any matching record into a counter; use a numeric OTTL expression for measured values.
• Leaving Resource Attributes empty passes through every incoming resource attribute, which can produce high-cardinality metrics. List only the keys you need.

Troubleshooting

No metrics are produced

Symptoms: the connector runs but emits nothing into the metrics pipeline.

Solutions:

1. Confirm the metric definition's Signal Type matches an enabled Telemetry Type and that the source pipeline is actually carrying that signal.
2. Check the Condition: if it never evaluates true for incoming records, no metric is generated. Temporarily remove it to confirm.

Output metric is missing expected attributes

Symptoms: datapoints appear but lack attributes you configured.

Solutions:

1. A Required attribute that is absent on an incoming record causes that whole record to be skipped. Switch to Optional or Has Default if the attribute is not always present.
2. Verify the attribute Key matches the source field exactly, including case.

Standalone Connector

apiVersion: bindplane.observiq.com/v1
kind: Connector
metadata:
  name: signal-to-metrics
spec:
  type: signaltometrics
  parameters:
    - name: telemetry_types
      value:
        - Logs
        - Traces
    - name: metrics
      value:
        - signalType: logs
          name: log.record.count
          type: sum
          value: "1"
          attributes:
            - key: severity
              behavior: optional
        - signalType: spans
          name: http.server.duration
          unit: ms
          type: histogram
          value: Double(attributes["http.server.request.duration"])
          condition: kind == SPAN_KIND_SERVER
          buckets:
            - 5
            - 10
            - 25
            - 50
            - 100
            - 250
            - 500
            - 1000
            - 2500
            - 5000
            - 10000
          attributes:
            - key: http.route
              behavior: optional
            - key: http.request.method
              behavior: required

Related Resources

• Signal to Metrics connector reference
• OTTL functions reference

Bindplane Resources

• Count
• Span Metrics