Skip to main content

Set Up Integration Management

Last updated on

Harness AI SRE connects to the tools you already use across monitoring, CI/CD, ticketing, communication, and on-call. Each integration feeds AI SRE the signals it needs to detect, investigate, and resolve incidents, or lets AI SRE act on those tools during a response.

What an integration gives you​

An AI SRE integration connects an external tool to AI SRE in one of two ways. A single tool can do both.

  • Ingest: The tool sends data into AI SRE. Monitoring tools send alerts that open or update incidents, and CI/CD tools send build, deploy, and pull request data the Deploy Change Investigator correlates to incidents. Ingest is one-way, into AI SRE.

  • Automation: AI SRE acts on the tool. Runbooks post messages, create tickets, page responders, and open bridges. Automation is one-way, out of AI SRE.

Ingest and automation are separate capabilities. A tool that only ingests, such as Datadog, sends alerts to AI SRE but cannot be acted on by a runbook. A tool that only automates, such as Slack, receives runbook actions but sends no incident signal. Tools such as Jira and ServiceNow do both.

Behind these capabilities are three transport mechanisms:

  • Native. Harness Code, Harness Pipelines, and Harness SLO integrate in-platform with no connector or webhook configuration required.

  • Webhooks. These are ingest-only and require no connector: the tool POSTs data to an AI SRE webhook URL. You create the webhook URL in AI SRE, then paste it into the sending tool. Go to Ingest Alerts to create a webhook.

  • Connectors. These call the tool's API and power automation, on-call schedule sync, and polling-based ingestion such as pull request and change-record ingestion. Configure connectors under Project Settings or Organization Settings → Third Party Integrations (AI SRE). Go to Integrate Your First Tool to connect one.

Where you manage integrations​

Open Integrations in the AI SRE left navigation to manage every connection to an external tool. The page has four tabs, each for a different kind of integration.

AI SRE Integrations page with the Integrations, Integration Templates, PR Ingestions, and Change Ingestions tabs in the top navigation

Use the tab that matches how the tool connects and what data it moves.

Integrations​

The Integrations tab lists your webhook endpoints for alert ingestion. Each entry has an endpoint URL that a monitoring tool POSTs alerts to, an Enabled toggle, and a Quiet Mode toggle that suppresses incident creation while you test the payload.

Integrations tab listing alert webhook endpoints, each with an endpoint URL, Enabled toggle, and Quiet Mode toggle

Use this tab when a tool sends alerts into AI SRE over a webhook, such as Datadog, Prometheus, or Grafana. Select New Integration to create an endpoint, then paste its URL into the sending tool. Go to Ingest Alerts to create a webhook.

Integration Templates​

The Integration Templates tab holds reusable field-mapping templates. A template maps an incoming payload to AI SRE incident and alert fields once, so several endpoints of the same type share one mapping instead of repeating it.

Integration Templates tab showing a Datadog Mappings template of type Alert

Use this tab when you ingest alerts from the same tool through more than one endpoint, or when you want a consistent mapping across a team. Go to Webhook Templates to review the prebuilt templates.

PR Ingestions​

The PR Ingestions tab lists pull request ingestion jobs. Each job connects a repository and syncs its merged pull requests so the Deploy Change Investigator can correlate a code change to an incident. The tab shows the repository, last sync time, and status, and lets you run a sync on demand.

PR Ingestions tab listing pull request ingestion jobs with repository, last sync time, and sync status

Use this tab to bring pull request and code-change data in from Harness Code, GitHub, or Bitbucket. Go to CI/CD and change sources to review the supported providers.

Change Ingestions​

The Change Ingestions tab lists connector-based change ingestion jobs. Each job uses a Harness connector to poll an external system, such as ServiceNow, for change records on a schedule. The tab shows the connector, last sync time, and status, and lets you retry a failed sync.

Change Ingestions tab listing connector-based change ingestion jobs with connector, last sync time, and status, including a failed job with a Retry action

Use this tab when change records reach AI SRE through a connector rather than a webhook, such as ServiceNow change requests. Go to ServiceNow to configure the connector.

Integration capabilities at a glance​

Use this table to confirm what AI SRE can do with each tool before you configure it. Ingest means the tool sends data to AI SRE. Automation means runbooks can act on the tool. Transport is how the connection is made. Where to manage is where you configure the tool: one of the four tabs on this page, Third-Party Integrations for automation connectors, or in-platform for native tools.

note

Configure automation connectors under Third-Party Integrations at either scope: Project Settings → Third Party Integrations (AI SRE) for a single project, or Organization Settings → Third Party Integrations (AI SRE) to share the connector across every project in the organization.

ToolIngestAutomationTransportWhere to manage
AlertSiteYesNoWebhookIntegrations tab
AWS CloudWatchYesNoWebhookIntegrations tab
BigPandaYesNoWebhookIntegrations tab
BitbucketYesNoConnectorPR Ingestions tab
CircleCIYesNoWebhookIntegrations tab
ConfluenceNoYesConnectorThird-Party Integrations
DatadogYesNoWebhookIntegrations tab
DynatraceYesNoWebhookIntegrations tab
GitHubYesYesConnector (pull requests, runbook actions), Webhook (build and deploy)PR Ingestions tab, Integrations tab, Third-Party Integrations
GitLabYesNoWebhookIntegrations tab
Google ChatNoYesConnectorThird-Party Integrations
GrafanaYesNoWebhookIntegrations tab
Grafana IncidentYesNoWebhookIntegrations tab
Harness CodeYesNoNativeNative (in-platform)
Harness PipelinesYesYesNativeNative (in-platform)
Harness SLOYesNoNativeNative (in-platform)
JenkinsYesNoWebhookIntegrations tab
JiraYesYesWebhook (deployments), Connector (runbook actions)Integrations tab, Third-Party Integrations
Jira Service ManagementNoYesConnectorThird-Party Integrations
LaceworkYesNoWebhookIntegrations tab
Microsoft TeamsNoYesConnectorThird-Party Integrations
New RelicYesNoWebhookIntegrations tab
Octopus DeployYesNoWebhookIntegrations tab
OpsGenieYesYesConnector (schedules), Webhook (alerts)Integrations tab, Third-Party Integrations
PagerDutyYesYesConnector (schedules), Webhook (alerts)Integrations tab, Third-Party Integrations
PrometheusYesNoWebhookIntegrations tab
SentryYesNoWebhookIntegrations tab
ServiceNowYesYesConnectorChange Ingestions tab, Third-Party Integrations
SlackNoYesConnectorThird-Party Integrations
SplunkYesNoWebhookIntegrations tab
TerraformYesNoWebhookIntegrations tab
Travis CIYesNoWebhookIntegrations tab
xMattersYesNoConnectorThird-Party Integrations
ZoomNoYesConnectorThird-Party Integrations

Monitoring and observability​

These tools ingest alerts into AI SRE to open and enrich incidents. They are receive-only: AI SRE does not act on them.

  • AlertSite: Ingest AlertSite monitoring alerts through a webhook.
  • AWS CloudWatch: Ingest CloudWatch alarms through an SNS webhook.
  • BigPanda: Ingest BigPanda alerts through a webhook.
  • Datadog: Ingest Datadog monitor alerts through a webhook.
  • Dynatrace: Ingest Dynatrace problem notifications through a webhook.
  • Grafana: Ingest Grafana alerts through a webhook.
  • Grafana Incident: Ingest Grafana Incident events through a webhook.
  • Harness SLO: Ingest Harness Service Reliability Management SLO alerts.
  • Lacework: Ingest Lacework security alerts through a webhook.
  • New Relic: Ingest New Relic alert conditions through a webhook.
  • Prometheus: Ingest Prometheus Alertmanager alerts through a webhook.
  • Sentry: Ingest Sentry issue alerts through a webhook.
  • Splunk: Ingest Splunk alerts through a webhook.

CI/CD and change sources​

These tools ingest build, deploy, and pull request data so the Deploy Change Investigator can tie a change to an incident. Most are receive-only. Harness Pipelines can also be triggered by runbooks, and GitHub also exposes runbook actions such as pull request creation.

  • Bitbucket: Ingest pull requests from Bitbucket Cloud.
  • CircleCI: Send build and deploy webhooks from CircleCI workflows.
  • GitHub: Ingest pull requests and send build and deploy webhooks.
  • GitLab: Send build and deploy webhooks from GitLab CI/CD.
  • Harness Code: Ingest pull requests automatically, with no connector required.
  • Harness Pipelines: Send build and deploy data, and trigger pipelines from runbooks.
  • Jenkins: Send build and deploy webhooks from Jenkins pipelines.
  • Octopus Deploy: Send deployment webhooks from Octopus Deploy.
  • Terraform: Send Terraform apply events as deployment data.
  • Travis CI: Send build and deploy webhooks from Travis CI.

Ticketing and ITSM​

These tools track change records and let runbooks create and update tickets during an incident. Jira and ServiceNow both ingest and automate; Jira Service Management and Confluence are automation-only.

  • Confluence: Create and update Confluence pages from runbooks.
  • Jira: Track Jira deployments and automate Jira ticket actions.
  • Jira Service Management: Create Jira Service Management alerts from runbooks.
  • ServiceNow: Ingest ServiceNow change records and automate ticket actions.

Communication​

These tools let your team collaborate on incidents and let runbooks post updates and open bridges automatically. They are automation-only: they carry no incident signal into AI SRE.

  • Google Chat: Collaborate on incidents in Google Chat spaces.
  • Microsoft Teams: Collaborate on incidents and post messages.
  • Slack: Collaborate on incidents and run Slack commands.
  • Zoom: Create meetings and conference bridges from runbooks.

On-call and escalation​

These tools sync on-call schedules into AI SRE and let runbooks page responders. PagerDuty and OpsGenie also run incident actions; xMatters is schedule sync only.

  • OpsGenie: Synchronize on-call schedules and create alerts.
  • PagerDuty: Synchronize on-call schedules and create incidents.
  • xMatters: Synchronize on-call schedules.

Custom integrations​

Connect a tool that is not listed above through a custom webhook:

  • Custom webhooks: Ingest alerts from any tool that can send an HTTP POST, using a custom webhook template with CEL or Mustache field mapping.

Next steps​