Delete Fields

The Delete Fields processor removes attributes, resource attributes, and (for logs) body fields from telemetry in the pipeline. It operates on logs, metrics, and traces. An optional OTTL condition restricts deletion to records that match; by default it applies to all telemetry of the selected types.

Supported Telemetry Types

Platform	Metrics	Logs	Traces
Linux	✓	✓	✓
Windows	✓	✓	✓
macOS	✓	✓	✓
AIX	✓	✓	✓
Kubernetes Node	✓	✓	✓
Kubernetes Cluster	✓	✓	✓
Kubernetes Gateway	✓	✓	✓
OpenShift Node	✓	✓	✓
OpenShift Cluster	✓	✓	✓
OpenShift Gateway	✓	✓	✓
Google SecOps Pipelines	✓	✓	✓

Configuration

General

Parameter	Type	Required	Default	Description
Choose Telemetry Type	Telemetry Selector	No	Logs, Metrics, Traces	The telemetry types this processor acts on.
Condition	OTTL Condition	No	(empty)	An OTTL condition to evaluate. Fields are deleted only from records that match. By default it applies to all telemetry.

Fields to delete

Parameter	Type	Required	Default	Description
Body Fields	OTTL Field	No	(empty)	List of body fields to delete. Body fields apply to logs only; shown when Logs is selected.
Attribute Fields	OTTL Field	No	(empty)	List of attribute fields to delete. Applies to logs, metrics, and traces.
Resource Fields	OTTL Field	No	(empty)	List of resource attribute fields to delete. Applies to logs, metrics, and traces.

Each field list is a repeatable row input. Add a row per field to delete. Use bracket notation for nested keys (for example data["region"]).

Examples

Remove a body field, an attribute, and a resource attribute from logs

This configuration deletes the spid body field, the log.file.name attribute, and the host.id resource attribute from every log record.

apiVersion: bindplane.observiq.com/v1
kind: Processor
metadata:
  name: delete-fields
spec:
  type: delete_fields_v2
  parameters:
    - name: telemetry_types
      value:
        - Logs
    - name: body_keys
      value:
        - spid
    - name: attributes
      value:
        - log.file.name
    - name: resource_attributes
      value:
        - host.id

Delete fields only from records that match a condition

This configuration deletes the http.request.header.authorization and http.request.header.cookie attributes from spans, but only where the service name is checkout. The underlying transform statement is delete_key(attributes, "http.request.header.authorization") where resource.attributes["service.name"] == "checkout".

apiVersion: bindplane.observiq.com/v1
kind: Processor
metadata:
  name: delete-fields
spec:
  type: delete_fields_v2
  parameters:
    - name: telemetry_types
      value:
        - Traces
    - name: condition
      value: resource.attributes["service.name"] == "checkout"
    - name: attributes
      value:
        - http.request.header.authorization
        - http.request.header.cookie

Configuration Tips

• Deleting metric attributes can be unsound. Removing an attribute may leave multiple data points with the same attribute set, causing data point collisions. Delete metric attributes only when you are sure the remaining attributes keep each series unique.
• Body fields apply to logs only. The Body Fields list is hidden unless Logs is one of the selected telemetry types, and it is ignored for metrics and traces.
• The processor uses the OTTL delete_key function, which targets one key at a time. Each field you list runs as a separate statement, so adding many fields per instance adds per-record work. To remove a whole family of keys by pattern, use the Transform (OTTL) processor with delete_matching_keys instead of listing each key here.
• Each enabled signal type renders its own transform processor, so a single instance selecting all three signals deploys three transform blocks to the collector.

Troubleshooting

Fields are not being removed

Symptoms: the target field is still present after the processor runs.

Solutions:

1. Confirm the field is listed under the correct group. Attributes, resource attributes, and body fields are distinct namespaces; a resource attribute listed under Attribute Fields will not match.
2. If a Condition is set, verify it evaluates true for the affected records. An overly narrow condition skips deletion.
3. Check the field name and nesting. Use bracket notation (for example metadata["host"]) for nested keys.

Body fields are ignored

Symptoms: a body field listed in the processor is never deleted.

Solution: Body fields apply to logs only. Ensure Logs is selected under Choose Telemetry Type. The Body Fields list has no effect on metrics or traces.

Metric series collapsed or counts look wrong after deletion

Symptoms: metric data points merge or values change unexpectedly after removing an attribute.

Solution: Deleting a metric attribute can make previously distinct series identical, causing data point collisions. Restore the attribute or delete a different one that is not part of the series identity.

Standalone Processor

apiVersion: bindplane.observiq.com/v1
kind: Processor
metadata:
  name: delete-fields
spec:
  type: delete_fields_v2
  parameters:
    - name: telemetry_types
      value:
        - Logs
        - Metrics
        - Traces
    - name: condition
      value: ""
    - name: body_keys
      value:
        - spid
    - name: attributes
      value:
        - log.file.name
        - http.request.header.authorization
    - name: resource_attributes
      value:
        - host.id

Related Resources

• Transform Processor — OpenTelemetry Collector Contrib
• delete_key and delete_matching_keys — OTTL function reference