Skip to main content

Configure On-Call Teams and Routing

Last updated on

Harness AI SRE uses User Groups from the Harness Platform as the organizational unit for on-call management. User Groups serve as teams in the on-call context: they own services, own escalation policies, and determine routing for alerts.

This page covers both administrator configuration (User Groups, service ownership, routing) and individual user configuration (personal notification settings).

User Groups as on-call teams​

Harness AI SRE does not have a separate concept of teams. Instead, it uses Harness User Groups defined at the platform level.

What are User Groups?​

User Groups are collections of users managed in the Harness Platform. They are used for:

  • Role-based access control (RBAC)
  • Service ownership in AI SRE
  • Escalation policy ownership in AI SRE
  • Alert routing in on-call management

Go to Add and Manage User Groups to create and manage User Groups in the Harness Platform.

User Groups in AI SRE on-call​

In the on-call context, User Groups determine:

  • Who owns which services: Each service in the Service Directory can be assigned an owning User Group.
  • Who owns escalation policies: Each escalation policy is owned by a User Group.
  • How alerts are routed: Alerts for a service route to the service's owning User Group's escalation policy.

Configure service ownership​

Services in the AI SRE Service Directory are assigned to User Groups. This determines which team is responsible for on-call coverage when alerts fire for that service.

Assign a User Group to a service​

Assign an owning User Group so alerts for the service route to the right team:

  1. Navigate to Project Settings → Service Directory (AI SRE).
  2. Find the service you want to configure.
  3. Click the service name to open its details.
  4. In the Owning User Group field, select the User Group responsible for this service.
  5. (Optional) Set an Escalation Policy Override if this service should use a different policy than the User Group's default.
  6. Click Save.

How service ownership works​

Service ownership drives alert routing through the following chain:

  • Each service has an owning User Group field.
  • When an alert fires for a service, AI SRE looks up the service's owning User Group.
  • The alert routes to that User Group's escalation policy.
  • The escalation policy determines which schedule (and therefore which on-call responder) receives the page.

Escalation policy ownership​

Escalation policies are owned by User Groups. This ties the policy to a specific team.

Policy owner field​

When creating or editing an escalation policy:

  • The Policy Owner field specifies which User Group owns this policy.
  • This typically represents the team that uses this escalation policy for their on-call rotation.
  • A User Group can have one default escalation policy, but you can create additional policies for specific scenarios.

Default escalation policy​

The default escalation policy determines routing behavior for a User Group:

  • Each User Group can have one default escalation policy.
  • When a service's owning User Group matches an escalation policy's owner, that policy is used for routing (unless the service specifies an override).
  • If a User Group has no default escalation policy, alerts for services owned by that User Group cannot be automatically routed.

Alert routing flow​

Alerts route through User Groups, services, and escalation policies as follows:

  1. Alert arrives with a service identifier in the payload.
  2. Service Directory lookup: AI SRE finds the service in the Service Directory.
  3. User Group resolution: AI SRE identifies the service's owning User Group.
  4. Escalation policy selection: AI SRE uses the service's escalation policy override (if configured), or the User Group's default escalation policy.
  5. Schedule lookup: The escalation policy references one or more schedules.
  6. On-call responder identification: AI SRE determines who is on-call at that moment.
  7. Notification dispatch: The on-call responder's personal notification rules are triggered.

This flow ensures alerts always reach the right person based on service ownership and team structure.

User notification settings​

While administrators configure User Groups and service ownership, each individual user configures their own contact methods and notification rules. These personal settings determine how you are notified when you are on-call and an incident is assigned to you.

Contact information​

Navigate to On-Call → Contact Settings to manage your notification channels.

Available contact methods​

You can configure the following notification channels:

  • Email: The email address tied to your Harness account. You can add additional email addresses (work, home, other) and send test notifications to verify delivery.
  • Phone: Add phone numbers with country code selection for voice calls, and send test calls to verify delivery.
  • SMS: Add mobile numbers for text messaging, and send test messages to verify delivery.
  • Slack: Link your Slack account to receive direct messages. This requires the Slack workspace integration to be configured by an admin, and supports one Slack account per user.
  • Mobile app: Install the Harness On-Call mobile app for push notifications. The app registers your device automatically when you log in. Download it from Google Play or the App Store.

Voice and SMS are supported in the following countries: United States (+1), Argentina (+54), Brazil (+55), Canada (+1), Costa Rica (+506), Cyprus (+357), Denmark (+45), India (+91), Israel (+972), Mexico (+52), Moldova (+373), Netherlands (+31), Romania (+40), Serbia (+381), Slovenia (+386), and United Kingdom (+44).

Add a contact method​

Add and verify a notification channel from Contact Settings:

  1. Navigate to Contact Settings in the On-Call menu.
  2. Click Add Email, Add Phone Number, Add SMS Number, or Add Slack.
  3. Enter the contact information.
  4. For phone or SMS, select the appropriate country code.
  5. Label the contact (Work, Home, Other).
  6. Click Save.
  7. Click Test to verify the contact method works.

Notification rules​

Notification rules define the sequence and timing of how you are notified when you are on-call. Each rule contains one or more steps, and each step specifies which channels to use and how long to wait before escalating.

Go to Configure Notification Fallback to see worked escalation sequences, create notification rule groups, understand per-channel retry behavior, and review the default notification sequence.

Troubleshooting​

Service alerts are not routing to the on-call responder in Harness AI SRE

Confirm the service has an owning User Group with a default escalation policy, the policy has active schedules, and the alert payload includes the correct service identifier.

An on-call user is not receiving notifications in Harness AI SRE

Verify contact methods and notification rules are configured, test each contact method, confirm the user is on-call at the time of the alert, and check the mobile app is logged in with notification permissions.

Slack notifications are not working in Harness AI SRE on-call

Confirm your organization's Slack workspace is connected to AI SRE, link your Slack account in Contact Settings, and test the connection. Only one Slack account can be linked per user.

Best practices​

For administrators​

Follow these practices when you configure User Groups and service ownership:

  • Align User Groups with actual team structure: User Groups should reflect real organizational teams.
  • Assign every production service to a User Group: Unmapped services cannot route alerts automatically.
  • Ensure every User Group has a default escalation policy: Without one, alerts cannot be routed.
  • Keep service ownership current: Update the Service Directory when teams change ownership.
  • Document naming conventions: Use consistent User Group names that clearly identify the team.

For on-call responders​

Follow these practices to make sure you are reached during your on-call shift:

  • Configure at least two contact methods: Redundancy ensures you are reached even if one channel fails.
  • Test your notification setup regularly: Use test buttons to verify delivery before your on-call shift.
  • Keep contact information current: Update phone numbers and email addresses immediately when they change.
  • Review your notification rules: Ensure escalation steps match your preferred notification sequence.
  • Install the mobile app: Push notifications are the most reliable way to receive urgent alerts.

Next steps​

Last updated on