Pipeline execution tags

Add tags dynamically to pipeline executions and filter by AND or OR logic.

Harness supports dynamic tagging of pipeline executions, enabling you to attach contextual metadata as tags during runtime. These tags become part of the execution record and can be used with Harness’s standard tag-based search and filters to quickly locate specific runs.

In real delivery workflows, critical information such as artifact versions, environment details, matrix dimensions, or dynamically generated configuration values often emerges only at runtime, not when the pipeline is authored. Dynamic tags allow you to capture this real-time context directly on the execution, giving you a richer, more accurate history of what was deployed, where, and under what conditions. This makes debugging, auditing, compliance checks, and operational insights significantly easier and more reliable.

---

What you will learn from this topic

• How to use expressions to add tags dynamically at runtime.

• How to understand limitations with shell script steps when adding tags.

• How to add tags with matrix looping for parallel executions.

• How to filter executions by tags using Matches All or Matches Any logic.

---

Before you begin

• Harness account access: You need access to a Harness account with pipelines. Go to Getting started with Harness Platform to create an account.

• Pipeline execution permissions: You need Execute permission on Pipelines to add tags to executions. To get this permission, an administrator must assign you a role that includes it. Go to RBAC in Harness to understand role-based access control and Manage roles to configure role assignments.

• Feature flag enabled: This feature requires the PL_EXECUTION_TAGS feature flag. Contact Harness Support to enable it.

---

Add tags dynamically to executions

You can add tags to pipeline executions dynamically at runtime using expressions to capture execution-specific information, such as artifact versions, environment details, or other values generated during the pipeline run. Dynamic tags are added only to the specific execution and can be used later to filter and find executions.

Expressions to add tags dynamically

The following expressions enable Harness to add tags dynamically to an execution:

Expression	Description
<+executionTags.addTag(name, value)>	attaches a single key-value tag to the execution at the moment the expression is evaluated.
<+executionTags.addTagsList(listOfPairs)>	attaches a list of tags to the execution at the moment the expression is evaluated.

Tags are not added to the pipeline definition. It exists only on the specific execution.

---

Limitations with shell script steps

These expressions cannot be invoked within a loop inside a Shell Script step. This is because all expressions are evaluated before the Shell Script step begins execution. Therefore, a loop in a script cannot call addTag multiple times and you cannot generate tag values inside the same script and immediately attach them.

If your workflow requires generating multiple tags in a script, use the following pattern instead:

• Generate a list of tags inside a script and export it as an output variable.

• Use <+executionTags.addTagsList(listOfPairs)> in the next step to attach them.

---

Add tags dynamically with matrix looping

You can use the expression <+executionTags.addTag(name, value)> with matrix looping strategy to add tags iteratively to an execution.

The step to add tags dynamically with matrix looping strategy will look like this:

The following sample pipeline YAML demonstrates how to add tags dynamically with matrix looping strategy:

Sample Pipeline YAML

pipeline:
  name: tags matrix
  identifier: tags_matrix
  tags: {}
  projectIdentifier: siddhant_playground
  orgIdentifier: default
  stages:
    - stage:
        name: Generate Tags
        identifier: shell
        description: ""
        type: Custom
        spec:
          execution:
            steps:
              - step:
                  type: ShellScript
                  name: ShellScript_1
                  identifier: ShellScript_1
                  spec:
                    shell: Bash
                    executionTarget: {}
                    source:
                      type: Inline
                      spec:
                        script: |-
                          export out="out"

                          json_output="{\"service\":["

                          # Generate 100 stack-branch combinations
                          for i in $(seq 1 10); do
                            stack="stack$i"
                            # Random-looking branch name using modulo logic
                            branch="branch$((i % 5 + 1))"  # branch1 to branch5
                            json_output+="{\"stack\":\"$stack\",\"branch\":\"$branch\"},"

                          done
                          # Remove the trailing comma and close the array
                          export json_output="${json_output%,}]}"

                          # Export as a string variable
                          echo "Exporting service output..."
                          echo "service=$json_output"
                          export service="$json_output"
                    environmentVariables: []
                    outputVariables:
                      - name: out
                        type: String
                        value: out
                      - name: service
                        type: String
                        value: service
                      - name: json_output
                        type: String
                        value: json_output
                  timeout: 10m
        tags: {}
    - stage:
        name: Add Tag
        identifier: poll
        description: ""
        type: Custom
        spec:
          execution:
            steps:
              - step:
                  type: ShellScript
                  name: ShellScript_1
                  identifier: ShellScript_1
                  spec:
                    shell: Bash
                    executionTarget: {}
                    source:
                      type: Inline
                      spec:
                        script: |
                          echo "stack is <+matrix.service.stack>"
                          echo "branch is <+matrix.service.branch>"
                          echo <+executionTags.addTag(<+matrix.service.stack>,<+matrix.service.branch>)>
                    environmentVariables: []
                    outputVariables: []
                  timeout: 10m
        tags: {}
        strategy:
          matrix:
            nodeName: <+matrix.service.stack>_<+matrix.service.branch>
            service: <+json.list("service", <+pipeline.stages.shell.spec.execution.steps.ShellScript_1.output.outputVariables.json_output>)>
            maxConcurrency: 100
        when:
          pipelineStatus: All

---

Filter pipeline executions by tags

When you view pipeline execution history, you can filter by multiple tags using AND or OR logic. This helps you narrow down results to find executions that match specific tag combinations.

Filter executions by tags

After you have added tags to your pipeline executions (either statically or dynamically), you can filter the execution history to find specific runs.

1. Go to Pipelines and select Execution History.

2. Select the Filters panel.

3. In the Pipeline Tags section, select the tags you want to filter by.

4. After selecting multiple tags, choose the operator under Conditions:

   • Matches Any: Returns executions that have at least one of the selected tags (OR logic)

   • Matches All: Returns only executions that have all of the selected tags (AND logic)

The default behavior is Matches Any (OR logic), which returns any pipeline that has at least one of your selected tags. When you switch to Matches All (AND logic), only pipelines that have every tag you selected will appear in the results.

Use cases for Matches All

The AND logic option (Matches All) is behind the feature flag PIPE_TAG_CONDITIONAL_FILTER. Contact Harness Support to enable the feature. OR logic (Matches Any) is available by default.

AND logic is most useful when you need to identify pipelines that meet multiple specific criteria simultaneously.

When troubleshooting production issues in a specific service, filter by tags such as env:prod AND service:payment-api to see only production deployments for that service, excluding test deployments or other services running in production.

For compliance audits, combine tags such as compliance:sox AND environment:production AND region:us-east-1 to generate reports showing only production deployments in SOX-compliant regions.

When multiple teams share a pipeline execution view, use tags such as team:platform AND priority:high to focus on high-priority work from a specific team.

Use cases for Matches Any

OR logic works well when you want to see pipelines matching any of several criteria.

When monitoring deployments across multiple regions, filter by tags such as region:us-east-1 OR region:eu-west-1 to see deployments in either region.

For reviewing work across related services, use tags such as service:payment-api OR service:checkout-service to view executions for any service in your payment flow.

Save filters with conditional logic

When you save a filter, Harness saves your tag filter operator choice along with the selected tags.

Backward compatibility

Filters created before this feature was enabled default to OR logic (Matches Any) when you edit them. If you want a saved filter to use AND logic, edit the filter and change the operator to Matches All, then save the filter again.

---

Next steps

• Assign metadata using tags: Learn how to structure tags and assign metadata in Harness.

• Add notes for pipeline execution: Add contextual notes to pipeline executions for better tracking.

• View and compare pipeline executions: View execution history and compare runs.