Skip to content
Tutorials Center
My learning

Topics

Lesson 6 of 38 ยท Flows and Triggers

Automated Cloud Flows and How Triggers Work

Learn how automated cloud flow triggers fire, the difference between polling and webhook triggers, common triggers for email, files and lists, and trigger settings.

An automated cloud flow starts when something happens in another service. Understanding how the platform notices that "something" explains most surprises: why a flow runs a minute late, why it runs twice, or why it never runs. This lesson covers the two trigger mechanisms and the triggers you will use most.

Polling and webhook triggers

There are two ways a trigger learns about an event.

Two ways to notice an event

Polling Power Automate asks every few minutes
Webhook The service tells Power Automate at once

Both look the same in the designer.

1 / 2
  • Polling triggers ask the service on a schedule, "anything new since last time?". The interval depends on your plan and the connector, commonly from one minute up to several minutes. A new item may wait for the next poll. Many SharePoint and Outlook triggers poll.
  • Webhook triggers register a callback. The service pushes a message the moment the event occurs. These are fast, and some Dataverse, Teams and HTTP triggers work this way.

The designer does not label which kind a trigger is, so check the connector's documentation if latency matters.

A poll that finds three new items starts three separate runs. A flow does not receive "a batch" unless the trigger is built for it, so design each run to handle one item.

Connector Trigger Fires when
Outlook When a new email arrives (V3) A message reaches a folder
SharePoint When an item is created / modified A list item is added or changed
SharePoint When a file is created in a folder A file appears
OneDrive When a file is created A file appears
Forms When a new response is submitted A form is answered
Teams When a new channel message is added Someone posts
Dataverse When a row is added, modified or deleted A table row changes

Many triggers are two-step: the trigger returns only an identifier (for example a Forms response id) and you then use a Get action (Get response details) to read the content.

Trigger settings

Open the three-dot menu of the trigger and choose Settings.

  • Split On is on for triggers that return a list: each item becomes its own run. Turn it off only if you really want to process the whole array inside one run.
  • Concurrency control limits how many runs happen at the same time. With it off, many runs start in parallel. With it on and a degree of 1, runs queue one after another, which protects shared resources such as a counter.
  • Trigger conditions filter items before a run starts (next lessons).
  • Networking, Security options restrict inputs for secure flows.

Duplicates and missed events

Why might a trigger run twice, or never?

  • Modified triggers fire for every change, including changes made by the flow itself. If your flow updates the item that triggered it, you can create an infinite loop. Prevent it with a trigger condition or a "processed" column.
  • A person edits an item three times quickly: with polling, you may get one run or three.
  • The trigger connection expired, so the flow is silently failing in the background. Check the flow's status page for an alert.
  • The flow was turned off, or the owner's license lapsed.

A stand-in: polling and Split On

This stand-in is plain Python. It simulates a poll that returns several new items and shows how "Split On" starts one run per item. It is not the real engine.

Python Stand-in: one poll, many runs (Split On)

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Set split_on = False and run it again to see what happens when the trigger hands over the whole array.

Avoiding the modified-trigger loop

Here is the loop problem in miniature. A flow reacts to a change and then changes the item, which triggers it again.

Python Stand-in: guard against an update loop

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Remove the processed check to see the loop. In a real flow, use a trigger condition such as @equals(triggerOutputs()?['body/Processed'], false), or write the update with a different account and filter on Modified By.

Best practices

  • Prefer triggers with filters (folder, subject, column) over filtering inside the flow.
  • Design a run to process exactly one item.
  • Keep a status column ("Pending", "Done") so the flow can be re-run safely.
  • Use a service account for flows that update the same list they listen to.

The next lessons cover buttons and clocks: Instant flows and manual triggers.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. What is a polling trigger?

    Polling triggers ask the service whether anything is new, so they may notice an event minutes later.
  2. 2. What does Split On do?

    Split On turns each element of the array returned by the trigger into its own run.
  3. 3. How can a 'when an item is modified' flow create an infinite loop?

    An update made by the flow is itself a modification, which fires the trigger again unless it is guarded.
  4. 4. What does concurrency control with a degree of 1 do?

    A degree of 1 queues runs so only one executes at a time.

Summary

  • Triggers learn about events by polling or by webhook; polling can add a delay.
  • Each item returned by a trigger normally starts its own run (Split On).
  • Concurrency control and trigger conditions limit and filter runs.
  • Guard "modified" flows against loops with a status column or a condition.

Next: Instant flows and manual triggers