Redesign your flow

  • 4 minutes

Occasionally, it's best to revisit the flow to determine if its design can be modified to improve the overall performance. This reassessment also allows you to apply new features or enhancements that weren't available when the flow was originally implemented.

How can Copilot help?

With the advent of Copilot, you can use the Copilot experience to also help identify Flow issues. For example:

Ask Copilot to review your flow by explaining its structure and purpose; or provide details like which triggers, actions, and loops are involved.

You can prompt Copilot to identify potential inefficiencies, like loops or conditions that could cause delays. Ask questions like:

  • "Are there any inefficient loops in this flow?"

  • "Can you find where the flow might be slowing down?"

When it comes to Optimization, you can also ask Copilot to suggest optimizations. Common areas to review include:

  • Triggers: Suggest using filters to avoid unnecessary triggers.

  • Loops: Check if you're using Apply to each efficiently and ensure that loops aren't processing unnecessary items.

  • Conditions: Identify where conditions could be simplified.

Ask Copilot questions like:

  • "Can you optimize my conditions?"

  • "Are there alternative actions I can use to make this flow faster?"

It can even help with Error Handling and Debugging. You can use Copilot to identify potential points of failure in your flow. Ask Copilot for help with analyzing run history and error messages:

  • "What errors are causing this flow to fail?"

  • "Can you check the error logs for a specific run?"

These and other prompts can help you identify issues, improve performance, and understand why your flows might be running slowly. However, having a set of eyes go through the process also help tremendously.

Use OData filter query to restrict returned entries

Because SharePoint connector is built on REST APIs, it supports your ability to filter data on the server side by using OData. Benefits of using OData include reducing the amount of data that you're bringing into your flow and reducing the need to loop through a record set to find values of interest.

Consider a list in Microsoft Lists that has over 100 items of computer device information. The list has a column called MFR Name that lists all manufacturers. The availability of this column and OData allows you to filter down to the manufacturer level on the server side, reducing the total amount of time that it takes for the flow to run.

Screenshot of a list of Devices.

In the SharePoint Get items action, in the Filter Query field, add the formula to filter the Microsoft devices.

Screenshot of the SharePoint Get items action.

Now, when you run the flow, only the Microsoft devices are visible.

Screenshot of the flow with the Your flow ran successfully message.

Modify the Do until condition

The Do until step does a certain action until a certain condition is true.

Changing the limits of the Do until condition can help speed the flow. The default setting is a count of 60, which is run every hour. Essentially, the flow checks every hour, 60 times, to determine if the condition has been met. The highest duration count is 30 days and a count of 5,000.

Screenshot of the Do until condition dialog box.

Approaches that you can take to improve performance:

  • Change the PT1H to PT24H for 24 hours, or change to PT72H, which means 72 hours.

  • Reduce the count, which will reduce the total number of loops.

  • Add a filter query if you're adding a query type action such as Get Items inside the Do until condition.

  • Leave the is successful check box selected in the Configure run after field. You don't want this step to run if the previous one has failed.

  • Avoid nesting of actions.

Reduce the frequency of scheduled flows

Power Automate allows you to create scheduled cloud flows that trigger a flow based on the starting time.

Screenshot of the New flow button selected and the Scheduled cloud flow option highlighted.

While you do have the flexibility to run flows every second, running them that often exhausts your API request limits.

Screenshot of the Build a scheduled cloud flow dialog box set to repeat every one minute.

All users of Microsoft Power Platform have limits on the number of requests based on the license that they're assigned. The following table defines the number of requests that a user can make in a 24-hour period.

User licenses Number of API requests per 24 hours
Power Apps per-user plan 5000
Power Automate per-user plan 5000
Microsoft 365 licenses 2000
Power Apps per-app plan 1,000 per app pass