Workflow Builder
Automate alert routing, notifications, and ticketing with a visual no-code workflow engine.
What is a workflow?
A workflow is an automation that runs when an alert or incident event occurs. You build it visually — drag nodes onto a canvas, connect them, and set conditions. No code required.
Each workflow has:
- A trigger — the event that starts it (e.g., alert firing, incident opened)
- Steps — actions, conditions, delays, and branches
- An execution log — every run recorded with inputs, outputs, and errors
Trigger types
| Trigger | Fires when… |
|---|---|
alert_firing | A new alert arrives with status firing |
alert_acknowledged | An alert is acknowledged |
alert_resolved | An alert's status changes to resolved |
alert_suppressed | An alert matches a suppression window |
incident_opened | A new incident is created |
incident_acknowledged | An incident is acknowledged |
incident_resolved | All alerts in an incident are resolved |
schedule | A cron schedule (e.g., "every day at 09:00") |
Available nodes
Branch based on alert fields — severity, source, client, labels, time of day.
Send SMS, voice call, email, Slack, Teams, or push notification.
Pause the workflow for N minutes before continuing.
Escalate the incident to the next on-call tier if still unacknowledged.
Open a ticket in Kaseya VSA or ConnectWise Manage.
Call any external API — PagerDuty, Jira, custom webhook.
Automatically acknowledge the alert.
Suppress the alert (useful for known-noise conditions).
Use the AI assistant to classify, summarize, or suggest resolution steps.
Execution model
Workflows execute asynchronously via a persistent job queue. Each execution:
- Is recorded with a unique run ID
- Shows the full step-by-step trace (inputs, outputs, duration)
- Is retried automatically on transient errors (network timeouts, rate limits)
- Fails permanently after 3 attempts, with the error logged
You can view all executions for a workflow at Workflows → [name] → Executions.
Limits
| Limit | Value |
|---|---|
| Steps per workflow | 50 |
| Nesting depth | 10 |
| Max delay per step | 24 hours |
| Concurrent executions per org | 100 |
| Execution history retention | 90 days |