Integrations for SRE
Integrations for SRE
Connect integrations in Workspace Settings -> Connections -> Integrations, then attach them to the SRE app or the agent that needs them. A connected tool provides access to its data; alert delivery is configured separately in Alerts -> Configure.
Follow the current integration setup guide for connection and attachment steps. The provider guides below describe credentials and capabilities using Platform 1.0 navigation; use the current Workspace Settings path when connecting them for SRE.
Choose context for the investigation
| Integration | Role in SRE | Setup reference |
|---|---|---|
| Grafana | Active alerts, alert rules, datasources, and available telemetry for discovery and investigation | Grafana credentials and capabilities |
| ObserveNow | Grafana-compatible alert delivery and observability context | Aiden for Observability |
| Datadog | Monitor alerts plus available metrics, logs, traces, and service context | Datadog credentials and capabilities |
| FireHydrant | Signals alerts and available incident, service, and runbook context | FireHydrant credentials and capabilities |
| GitHub | Repository files, configuration, workflow failures, and change context during investigation | GitHub credentials and capabilities |
| PagerDuty | Incident notifications through webhooks | PagerDuty setup |
| New Relic | Applied Intelligence issues and connected observability context | New Relic setup |
GitHub supplies investigation context; it is not an alert-ingestion provider in Configure Alerts. Likewise, an integration appearing in the catalog does not mean it supports every SRE delivery method. See the alert delivery matrix and Supported Integrations.
Connect and verify
- Select the workspace that will own the alerts and investigations.
- Add the integration and complete its credential fields. Use an instance name that identifies the account or environment.
- Attach the connection to SRE or its relevant agent. Check that the expected instance appears in Run Discovery or Configure Alerts.
- Run Discovery and review its history and knowledge report.
- Configure alert ingestion, then confirm that an alert's Source identifies the expected instance.
- Investigate one representative alert and verify that its evidence comes from the intended environment.
Multiple instances of one provider are configured separately. Check the instance name before changing a webhook, schedule, or filter.
Private networks and runbooks
If Discovery or Investigations shows Install a remote runner for in-VPC investigations, use Set up runner and the Remote Runners guide. The runner must be attached to the SRE investigator agent. A runner supplies a place to execute supported commands inside your network; credentials and permissions still determine the resources those commands can access.
Add operational guidance through Knowledge Hub and attach the relevant knowledge to the agent. Include service ownership, known failure modes, verification steps, and rollback instructions. For repository access, see the GitHub credential guide.
If an integration is missing
- Check the selected workspace and that the integration is attached to the app or agent.
- Check the delivery matrix: a context integration such as GitHub will not appear as an alert source.
- Verify credentials, provider permissions, endpoint reachability, and runner attachment where needed.
- If discovery succeeds but an investigation cannot query telemetry, inspect the investigation events for the failing tool or datasource. Discovery coverage does not prove every query is permitted.