Integration Steps Reference

Reference for integration step templates in Harness 3.0; HTTP, email, Jira, ServiceNow, Jenkins, Bamboo, SSH, WinRM, and AI Verify.

Harness 3.0 provides integration steps for HTTP requests, notifications, ticketing systems, CI tools, remote execution, and AI-based verification. These steps enable pipelines to interact with external systems and services.

HTTP step

Template: httpStep@1.0.0 · Module: CD/Custom

Execute HTTP requests with support for all methods, custom headers, mTLS authentication, response validation using CEL expressions, and output variable extraction. Runs in harnessdev/harness-http:0.0.1.

Input	Type	Required	Description
url	string	Yes	URL for the HTTP request
method	select	Yes (default: GET)	HTTP method (GET, POST, PUT, DELETE, PATCH, HEAD, OPTIONS, TRACE, CONNECT)
headers	string (textarea)	No	HTTP headers as JSON (e.g., {"Content-Type":"application/json"})
body	string (textarea)	No	Request body content
body_file	string	No	Path to file containing body
work_dir	string	No	Working directory (default: workspace)
assertion	string (textarea)	No	CEL expression for validation (e.g., response.code == 200)
output_vars	string (textarea)	No	JSON mapping for output variables
env_vars	key-value-pairs	No	Environment variables
client_cert	string (secret)	No	Client certificate for mTLS
client_key	string (secret)	No	Private key for mTLS
skip_verify	boolean	No (default: false)	Skip SSL/TLS verification
proxy	string	No	HTTP/HTTPS proxy URL
disable_redirect	boolean	No (default: false)	Disable redirect following
log_level	select	No (default: debug)	Log level

CEL EXPRESSION SUPPORT

The HTTP step uses CEL (Common Expression Language) for response assertions and output variable extraction. Use response.code for status code, response.body for parsed JSON body, and response.headers for response headers.

steps:
  - name: Health Check
    uses: httpStep@1.0.0
    with:
      url: https://api.example.com/health
      method: GET
      assertion: response.code == 200

steps:
  - name: Create Resource
    uses: httpStep@1.0.0
    with:
      url: https://api.example.com/resources
      method: POST
      headers: '{"Content-Type":"application/json","Authorization":"Bearer <+secrets.getValue(\"api_token\")>"}'
      body: '{"name":"new-resource","type":"production"}'
      assertion: response.code == 201
      output_vars: '{"resource_id":"response.body.id"}'

steps:
  - name: Secure API Call
    uses: httpStep@1.0.0
    with:
      url: https://internal-api.example.com/data
      method: GET
      client_cert: <+secrets.getValue("client_cert")>
      client_key: <+secrets.getValue("client_key")>

---

Email step

Template: email@1.0.0 · Module: CD

Send email notifications to recipients and user groups. Runs in harnessdev/email:0.0.1.

Input	Type	Required	Description
email_ids	string	Yes	Primary recipient email addresses
subject	string	Yes	Email subject line
body	string (textarea)	Yes	Email message content
cc_email_ids	string	No	CC email addresses
to_user_groups	string	No	Primary recipient user groups
cc_user_groups	string	No	CC user groups
smtp_config	string	No	SMTP configuration
attachments	list	No	Email attachments
log_level	select	No (default: info)	Log level

steps:
  - name: Notify Team
    uses: email@1.0.0
    with:
      email_ids: "[e-mail],[e-mail]"
      subject: "Deployment Complete - <+pipeline.name>"
      body: |
        Pipeline <+pipeline.name> has completed successfully.
        Environment: <+env.name>
        Build: <+pipeline.sequenceId>

---

Jira integration steps

Three integration steps for creating, updating, and approving via Jira issues.

Step	Template ID	Version	Description
Jira Create	jiraCreate	1.0.0	Create a new Jira issue
Jira Update	jiraUpdate	1.0.0	Update an existing Jira issue
Jira Approval	jiraApproval	1.0.0	Approve or reject via Jira issue

Key inputs (create step):

Input	Type	Required	Description
connector	connector (Jira)	Yes	Jira connector
project	string	Yes	Jira project key (e.g., PROJ)
issue_type	string	Yes	Issue type (Story, Bug, Task, etc.)
fields	key-value-pairs	No	Additional fields to set

steps:
  - name: Create Bug Ticket
    uses: jiraCreate@1.0.0
    with:
      connector: account.jira
      project: PROJ
      issue_type: Bug
      fields:
        summary: "Deployment failed - <+pipeline.name>"
        description: "Pipeline execution <+pipeline.executionId> failed."
        priority: High

steps:
  - name: Jira Approval
    uses: jiraApproval@1.0.0
    with:
      connector: account.jira
      issue_key: <+pipeline.variables.jira_ticket>
      approval_criteria:
        status: Approved

---

ServiceNow integration steps

Four integration steps for creating, updating, approving, and importing records via ServiceNow.

Step	Template ID	Version	Description
ServiceNow Create	serviceNowCreate	1.0.0	Create a ServiceNow ticket
ServiceNow Update	serviceNowUpdate	1.0.0	Update an existing ticket
ServiceNow Approval	serviceNowApproval	1.0.0	Approve or reject via ServiceNow
ServiceNow Import Set	serviceNowImportSet	1.0.0	Import a set of records

steps:
  - name: Create Change Request
    uses: serviceNowCreate@1.0.0
    with:
      connector: account.servicenow
      table: change_request
      fields:
        short_description: "Deploy <+service.name> to production"
        category: Software
        priority: "2"

---

Jenkins build

Template: jenkinsBuild@1.0.0 · Module: CI

Execute a Jenkins job and monitor its progress.

Input	Type	Required	Description
connector	connector (Jenkins)	Yes	Jenkins connector
job	string	Yes	Jenkins job name/path
params	key-value-pairs	No	Job parameters

steps:
  - name: Trigger Jenkins Build
    uses: jenkinsBuild@1.0.0
    with:
      connector: account.jenkins
      job: my-project/main
      params:
        BRANCH: main
        DEPLOY_ENV: staging

---

Bamboo build

Template: bambooBuild@1.0.0 · Module: CI

Execute a Bamboo plan and monitor its progress.

Input	Type	Required	Description
connector	connector (Bamboo)	Yes	Bamboo connector
plan	string	Yes	Bamboo plan key
params	key-value-pairs	No	Build parameters

steps:
  - name: Trigger Bamboo Build
    uses: bambooBuild@1.0.0
    with:
      connector: account.bamboo
      plan: PROJ-PLAN

---

SSH step

Template: ssh@1.0.0 · Module: CD

Execute a shell script on a remote host using SSH. Runs in harnessdev/ssh:0.0.1.

Input	Type	Required	Description
host	string	Yes	Target SSH host/IP
script	string	Yes	Shell script to execute
port	integer	No (default: 22)	SSH port
work_dir	string	No	Remote working directory
env_vars	key-value-pairs	No	Environment variables
out_vars	key-value-pairs	No	Output variables to capture
secret_out_vars	key-value-pairs	No	Secret output variables
script_timeout	string	No (default: 30m)	Execution timeout
log_level	select	No (default: info)	Log level

steps:
  - name: Deploy to Server
    uses: ssh@1.0.0
    with:
      host: 10.0.1.100
      port: 22
      script: |
        cd /opt/app
        docker pull myorg/myapp:latest
        docker-compose up -d
      env_vars:
        APP_VERSION: <+pipeline.sequenceId>
      out_vars:
        deploy_status: DEPLOY_STATUS
      script_timeout: 10m

---

WinRM step

Template: winrm@1.0.0 · Module: CD

Execute a PowerShell script on a remote Windows host using WinRM.

Input	Type	Required	Description
host	string	Yes	Target WinRM host
script	string	Yes	PowerShell script to execute
port	integer	No (default: 5985)	WinRM port
env_vars	key-value-pairs	No	Environment variables
out_vars	key-value-pairs	No	Output variables
script_timeout	string	No (default: 30m)	Timeout

steps:
  - name: Deploy to Windows
    uses: winrm@1.0.0
    with:
      host: 10.0.1.200
      script: |
        Stop-Service -Name "MyApp"
        Copy-Item "\\share\builds\latest\*" "C:\MyApp\" -Recurse
        Start-Service -Name "MyApp"
      script_timeout: 15m

---

AI Verify step

Template: aiVerifyStep@2.0.0 · Module: CV

AI-based verification for continuous deployment with intelligent anomaly detection and log analysis. Supports multiple health source types including Datadog Metrics, Dynatrace Grail Logs, and generic log sources. Uses a matrix strategy to run verification across all configured health sources.

Input	Type	Required	Description
health_sources	array	Yes	Health source identifiers to verify
duration	string	Yes (default: 5)	Verification duration in minutes
sensitivity	select	Yes (default: Medium)	Anomaly detection sensitivity (Low, Medium, High)
fail_no_data	boolean	No (default: true)	Fail if no analysis data
fail_metric	boolean	No (default: true)	Fail if any metric has no analysis
control_regex	string	No	Regex for control nodes (e.g., .*primary.*)
test_regex	string	No	Regex for test nodes (e.g., .*canary.*)
baseline_end	string	Yes	End time of pre-deployment analysis window
verify_start	string	Yes	Start time of post-deployment analysis window
context	string	No	Natural language context for monitoring
service	string	No	Service identifier (default: from context)
env	string	No	Environment identifier
infra	string	No	Infrastructure identifier

Verification workflow: (1) Maps health sources to appropriate data collection plugins (Datadog, Dynatrace, etc.). (2) Runs data collection in parallel using matrix strategy across health sources. (3) Compares baseline (pre-deployment) metrics with verification (post-deployment) metrics. (4) Uses AI to detect anomalies and analyze logs. (5) Reports pass/fail based on sensitivity configuration.

steps:
  - name: AI Verification
    uses: aiVerifyStep@2.0.0
    with:
      health_sources:
        - datadog_metrics_prod
        - dynatrace_logs_prod
      duration: "10"
      sensitivity: Medium
      context: "Verify new deployment has no performance regression"

HEALTH SOURCE PLUGIN SELECTION

AI Verify automatically selects the appropriate data collection plugin based on your health source type. Datadog Metrics uses the metrics plugin, Dynatrace Grail uses the log plugin, and other sources use the default log analysis plugin.