Why did I stop using PowerFlow?
Two weeks ago, I found myself enamoured with the vast capabilities of the Power platform. Admittedly, my familiarity was limited to PowerBI and PowerQuery. PowerApp and PowerFlow were unexplored territories.
A chance encounter revealed the potential of PowerApp to address critical business challenges. Many businesses, reliant on the MS Office suite, were drowning in unconnected Excel spreadsheets, creating data silos that hindered the flow of information.
PowerFlow is characterized as the low code tool to automate the business flow of an organisation. No one knows better than the business logic owner about how the business is operated. With the PowerFlow, orchestrated business flow can siege the information together, pipelined, and aggregated. What a brilliant idea!
I blamed myself for not being able to know the Powre platform earlier. It is a treasure, and I decided to do some digging. I decided to build an MVP using PowerFlow and PowerApp.
My MVP aimed to extract deal information from a CRM SaaS through REST APIs and populate an Excel workbook—a typical software prototype. The initial stages of building the PowerFlow felt like drawing a software flow chart—swift and intuitive. However, during testing, a hiccup emerged.
领英推荐
Take a closer look at these parallel running branches. The "Apply to each" step takes significant time to complete. "Apply to each" is basically a for-loop control. In this case, the iteration for around one hundred times takes more than 10 seconds! The most complicated operation is to append an item to an array and avoid duplication. I/O operation is normally a bottleneck of classic software design. However, the REST API calls are completed like a flash compared to the for loop.
I stretch my hair and believe that I must do something wrong. I look for the answer online and find that I'm not alone. Many PowerFlower users complain about the "Apply to each" performance issue. Some users wait over 40 hours to insert 5000 rows into the SharePoint list. The issue was reported as early as 2020. There has been no software behaviour update in the past three years.
I was shocked! How could a tool designed to improve productivity and eliminate data silos fall short in performance? So my MVP hit the stop sign, and I stopped probing.
Admitting my newcomer status to PowerFlow, I realized that my usage pattern of "Apply to each" might not be an ideal one. Some engineers had outlined best practices to sidestep this performance trap, but the absence of a traditional "for loop" posed a challenge. How could I convince clients to embrace a low-code solution with such limitations?
My PowerFlow journey has indeed been a rollercoaster. While the Power platform is undoubtedly a great product, my anticipation is anchored in the hope that Microsoft will attentively heed user feedback and bring about substantial improvements.