Managing Multiple Actions Execution in Flow: A Comprehensive Guide

Main Subject

Power Automate's processing performance could be challenging when a massive amount of actions are executed simultaneously. One apply this in the context where numerous rows are processed, or operations are conducted quickly.

Identifying an instance from a personal blog was an exemplary workflow that sends feedback requests from employees to their main collaborators. A system like this functions efficiently in a small-scale organization say with 50 members. However, the situation might change when the organization's scale is magnified for example up to 2,500 employees.

The first visible limitation surfaced in the process of distributing adaptive cards. The workflow must wait for a user response before moving forward to further business logic. To rectify this, we can utilize the degree of parallelism of the outer loop.

Secondary Explorations

Adapting child flows is beneficial to complex workflows. Among their usages are dividing intricate workflow into several separate workflows, increasing parallel runs, easier maintenance, and initiating the same flow from multiple distinct flows.

A child flow is used to manage the collection of one person's feedback, namely, identifying recipients, circulating feedback requests, and if given, saving responses. Child flows must return a response to the caller; thus, responses are returned instantly on the parallel branch. Responses are implemented using premade connections; however, the connector's capacity restricts it.

Feedback items retrieved from various sources experience paging limitations which could be resolved by turning on Pagination and determining the processing line's maximum number.

Challenge Areas

Two vital elements to be aware of are the connection throttling and Power Platform requests. Depending on the Teams operations count in the child flow, it could only launch a certain amount due to the connector's limitations. Adding delay to the main flow can help address this challenge.

With regards to Power Platform requests, the child flow execution generates a specific number of requests, and additional requests are executed over a longer period. The main workflow consumes a significant number of requests as well.

Summary

A successful navigation of Power Automate when dealing with thousands of data rows could present unfamiliar hurdles. It could at times feel slow from a developer's perspective, yet fast enough that restrictions related to connections usage are easily surpassed.

Understanding these challenges primarily surface in real-world scenarios and not during testing phases with smaller data sets is crucial. Therefore, three main points should be considered when creating Power Automate flows:

• What is the value of Degree of Parallelism?
• How will Connection Throttling impact performance?
• How many Power Platform requests will be consumed?

For more information on Power Automate, visit this link.

Read the full article How does flow cope when a lot of actions are excecuted?

Learn about How does flow cope when a lot of actions are excecuted?

Understanding Power Automate, previously known as Microsoft Flow, requires a thorough examination of its functions, especially when many actions are executed rapidly. One might question how this software manages in situations such as those.

In an analysis, we will refer to a scenario previously presented in a personal blog that discussed a particular workflow process. Specifically, this process sends out feedback requests from employees to their primary colleagues.

The system functions well in a small organization with approximately 50 people. However, the real challenge manifests when used in an organization with upwards of 2,500 employees. Essentially, the software struggles to efficiently manage vast amounts of information and operations.

The initial hurdle in this flow is the act of sending an adaptive card. The process does not continue until the user has responded to the adaptive card, creating a bottleneck. This issue can be resolved by adjusting the degree of parallelism of the outer loop (Apply to each) to full (50), allowing 50 feedback requests to occur simultaneously.

A more substantial solution to the problem lies in using child flows, which allows for the division of a complex flow into several flows. A child flow handles the collection of feedback from one person and saves replies, if there are any. However, a problem arises due to the Microsoft Teams action used in our child flow: each connector has its limitations.

To address this issue, specific delays can be added in the main flow after the child flow has started. Though this method slows down the process, it ensures that the limits of the connectors are not exceeded. If we were to apply this to an organization with 2,500 employees, it would take more than 40 hours to gather feedback from everyone, consuming 2,500 additional requests in the process.

In terms of licensing, an Office 365 account can execute 6,000 requests per 24 hours. Alternatively, a Power Automate per-user license allows for 40,000 requests per 24h, enough for our requirements as the flow will take around 40 hours to complete. Lastly, a flow license allows for up to 250,000 requests per day, providing more than enough capacity for our scenario, barring any extreme cases.

Through this scenario, it becomes clear that using Power Automate cloud flow to manage thousands of data rows can present new and unfamiliar challenges. From a developer's perspective, the software can often be slow, yet simultaneously fast enough to easily surpass the restrictions related to the use of connections. These problems seldom arise when testing the flow with a small amount of test data.

When constructing new flows, some primary considerations should include synchronizing operations, allocating resources efficiently, and managing user responses efficiently. These factors can significantly influence the success of your automated processes.

Expanded learning opportunities are available through entities such as the Forward Forever monthly newsletter, which provides further insights and articles regarding Power Apps, Power Automate, Power BI, among others.

More links on about How does flow cope when a lot of actions are excecuted?

How does flow cope when a lot of actions are excecuted?

Dec 5, 2022 — Child flow – Collect feedback from one person · divide a complex flow into several separate flows that call each other · trigger the same flow ...

Using asynchronous actions in a Flow

Feb 24, 2022 — Salesforce recently added another Flow ability: Asynchronous actions. This post discusses how to use asynchronous actions in Flows.

What Happens When a Flow Finishes?

By default, when a flow interview that uses screens finishes, a new interview for that flow begins, and the user is redirected to the first screen. To ove.

Keywords

Flow execution, cope with actions, action overload, process management, efficient workflow, handling multiple actions, flow performance, action execution, workflow optimization, process execution.