Start Node

Configure the Start node to define when and how your workflow is triggered.

  3 minute read  

Overview

The Start node is the entry point for every workflow. It defines when the workflow executes and what data is available to downstream nodes. Every workflow must have exactly one Start node.

Start node configuration panel Start node configuration panel

Trigger types

Configure the Start node to respond to one or more trigger types.

Monitors

Run the workflow when a monitor fires an alert.

To connect a monitor:

  1. Click Add Monitor to see the connection instructions
  2. Navigate to the monitor you want to connect
  3. In the monitor’s notifications section, type @ and select this workflow

The monitor’s alert data becomes available as variables for downstream nodes.

Connectors

Run the workflow when events arrive from a configured connector.

Select the connector and event type that should trigger the workflow. The available event types depend on which connectors are configured in your organization. See Connectors for setup instructions.

Example connector events:

ConnectorExample events
GitHubPR created, PR merged, issue opened, deployment status changed
PagerDutyIncident triggered, incident resolved, incident acknowledged
SlackMessage received, reaction added
CircleCIBuild completed, build failed
SentryIssue created, issue resolved
LinearIssue created, issue updated
LaunchDarklyFlag changed
JenkinsBuild completed
GitLabPipeline completed, merge request created

The connector event payload becomes available as variables for downstream nodes.

Periodic run

Run the workflow on a recurring schedule.

Configure the schedule using a cron expression. Examples:

ScheduleCron expression
Every day at 9:00 AM0 9 * * *
Every 6 hours0 */6 * * *
Every Monday at 8:00 AM0 8 * * 1
Every 15 minutes*/15 * * * *

Periodic workflows are useful for:

  • Scheduled health checks and infrastructure reviews
  • Daily, weekly, or monthly report generation
  • Compliance audits and security posture checks
  • Capacity planning and cost monitoring
  • AI-assisted reviews (pair with a Start AI Conversation action to create interactive review threads; see the Periodic AI review pattern)

For a working example, the Default Company Page Daily Summary workflow runs daily at 9:00 AM to check synthetic monitor health and email results.

Best practices for periodic workflows

Choose the right frequency. Start conservatively and increase frequency if findings are actionable. High-frequency schedules (every 15 minutes) suit critical systems requiring rapid detection. Daily schedules work well for status summaries and trend monitoring. Weekly or monthly schedules suit compliance reviews and aggregated reports.

Design actionable prompts. Define clear success criteria in the teammate prompt (“no pods in CrashLoopBackOff” rather than “check Kubernetes”). Specify what constitutes a finding worth reporting and include context that helps prioritize results. If a periodic workflow generates noise without driving decisions, adjust the prompt or reduce frequency.

Pair with If/Else for conditional actions. Use an If/Else node after the teammate analysis to branch on findings. For example, only send an email when the teammate detects issues, and skip notifications when everything is healthy.

Manual execution

You can trigger any workflow manually using the Run Manual button in the workflow editor toolbar. This is useful for testing workflows or running ad-hoc automation.

Viewing run sources

The workflow list displays the run source for each execution, showing whether it was triggered:

  • Manually
  • By a periodic schedule
  • By a monitor event
  • By a connector event