21.5 Integration Service Event Triggers: Connected and Disconnected

21.5 Integration Service Event Triggers: Connected and Disconnected

Core Concept: Event triggers start automations when something happens in another application, such as an email arriving in Outlook or a record being created in Salesforce. Integration Service detects these events, and Orchestrator starts the job. The exam description lists event triggers under API automation.

---

Anatomy of an Event Trigger

The available filter values depend on the selected connection.

---

How Events Are Detected: Polling

Connectors support events through a polling mechanism:

• Integration Service queries the application's API at an interval, every five minutes by default.
• The polling interval is set at the connection level, so changing it affects all triggers on that connection.
• Expect a delay of up to the polling interval between the event and the job start.

In Integration Service, the only trigger setting you change is this polling interval; the triggers themselves are managed in Orchestrator.

---

Connected Versus Disconnected Triggers

Arguments for disconnected triggers

A process started by a disconnected trigger receives four String arguments:

Declare these as In arguments of type String. Use UiPathEventObjectId with a Get Record or HTTP activity to fetch the full record.

---

Permissions

Creating or editing an event trigger in Orchestrator needs:

• Connections View, and assignment to the folder where the connection resides.
• Resource overwrites permissions (View, Edit, Create, Delete) to use or configure a connection other than the one originally in the workflow, including per-user configurations.

---

Designing Triggered Processes

1. Filter early. Use the trigger's event filters so jobs start only for relevant events.
2. Keep the process idempotent. Polling and retries can mean a record is seen more than once; check whether the work was already done.
3. Fetch fresh data. Especially for disconnected triggers, read the record with its ID at the start of the job.
4. Log the event identifiers for troubleshooting.
5. Check the attempts history. Integration Service keeps an attempts history for each trigger, which shows whether an event was detected even when no job ran.

---

Event Triggers Versus Other Triggers

Use an API trigger when the other system can call UiPath directly and you need an immediate start; use an event trigger when you only need to react to changes UiPath can detect through the connector.

---

Key Terms

• Polling interval: how often Integration Service checks the application for new events; set per connection.
• Event filter: conditions on the event, such as a mailbox folder or record type, that decide whether a job starts.
• Package requirement: a connection or trigger a process needs, shown when the process is added in Orchestrator.
• Resource overwrite: a replacement connection chosen for a process or user instead of the one used at design time.

---

Troubleshooting Event Triggers

• The job starts minutes after the event: this is expected with polling; lower the connection's polling interval if the delay matters, remembering that it applies to every trigger on that connection.
• No job started: check the trigger's attempts history, the event filters, and whether the connection is healthy or needs to be reconnected.
• A disconnected trigger's process fails immediately: confirm the process declares the four UiPathEvent... arguments as String In arguments with the exact names.
• Wrong mailbox or account in production: check the connection selected through Package Requirements or the resource overwrite for the process.

---

Worked Example: Supplier Invoices by Email

1. A connected trigger "Email received" on the AP team's Outlook connection, filtered to the Invoices folder and emails with attachments, starts the process.
2. The process saves the PDF attachments to a storage bucket and adds a queue item per invoice.
3. In test and production, the process uses different mailboxes; the connection is chosen per environment through Package Requirements.
4. The team lowers the connection's polling interval because invoices should be picked up faster, knowing this affects every trigger on that connection.