Has anyone come up with any easily quantifiable methods of testing the performance of the FME Flow user interface?
I’ve sat with a stopwatch in the past but it’s a bit of a laborious process.
Has anyone come up with any easily quantifiable methods of testing the performance of the FME Flow user interface?
I’ve sat with a stopwatch in the past but it’s a bit of a laborious process.
Doesn’t the CPU time indication at the end of the log suffice? What’s your use case?
Or do you mean the GUI itself, rather than the workspace execution.
GUI rather than workspace execution.
I’ve come across issues previously where excessive logging on the FME core database led to an extremely slow user interface (slow enough that the stopwatch was measuring in minutes)
Other times, performance between different environments seems to vary somewhat but it’s hard to define whether that’s real or perception.
I’d like to be able to monitor whether there’s any degradation in performance or whether performance is similar to existing in a new environment
Perhaps you can use the Performance tab of the browser developer tools, I believe both Chrome and Firefox have this functionality.
David is right: Press F12 in (as far as I know) all modern browsers (but at least in Chrome, FireFox, and Edge - those are the ones I use on Windows, Opera I use only on Android). That will open the developer tools, and in there (sometimes called performance, sometimes called network) you can find performance of the currently open page - although you have to reload if you opened it with the page you want to monitor already open.
It’s quite funny: back in the day google was pushing “let’s make the web faster”, so when I used the network tools in Chrome (!) on the 2 public websites I was maintaining at the time, it consistently showed that the stuff like analytics and translations which I loaded from google were the slowest to load But those tools are quite invaluable for things like this.
Most calls in the FME Server UI are just rest calls. If you’re able to identify one that is slow (using the dev tools as suggested) then you should be able to do a somewhat direct comparison between environments using something like FME or Post Man on just that specific call, this does assume that no other things are in the way (e..g, proxies) or nothing has previously been cached.